AI News Archive: July 8, 2026 — Part 5
Sourced from 500+ daily AI sources, scored by relevance.
- NEWSLETTER: China weighs silicon curtain around sought-after AI models
NEWSLETTER: China weighs silicon curtain around sought-after AI models Reuters
- Why AI provider dependency is becoming Europe’s next startup risk (Sponsored)
A single provider decision can now shape the future of an AI startup. For many European founders, that sounds dramatic until the product is already in customers’ hands. A model update changes the quality of an answer, pricing moves just as usage begins to climb, access rules tighten, and a security team asks where customer […] The post Why AI provider dependency is becoming Europe’s next startup risk (Sponsored) appeared first on EU-Startups .
- Slack’s Slackbot can now pull your CRM data, generate charts, and send DocuSigns — all from a chat message.
Five years and $27.7 billion after Salesforce acquired Slack, the two products are finally starting to function as a single system. On Wednesday, Slack launched an integration that connects Slackbot — the personal AI agent built into every workspace — to the entire Salesforce platform, including CRM data, Tableau analytics, Data 360 customer profiles, and a growing constellation of third-party applications, all through a single conversational prompt. The mechanism behind the expansion is a set of dedicated Model Context Protocol (MCP) servers from Salesforce that connect Slackbot to the company's Headless 360 infrastructure . In practical terms, a salesperson can now ask Slackbot for a customer's deal history, receive a live Tableau visualization of pipeline trends, update a CRM record, and trigger a DocuSign approval — without ever switching tabs or logging into another application. According to Slack, the Salesforce IT team has already used this architecture to save its 1,500-plus engineers "thousands of custom coding hours annually." The timing is not accidental. Slack is making this move amid escalating competitive pressure from Microsoft Teams, which claims 320 million-plus monthly active users and has Copilot embedded across the Office suite, and from Google, which continues to weave Gemini deeper into Workspace . And just days ago, The Information reported that some smaller companies are using Anthropic's Claude to r eplace Salesforce CRM entirely — one Atlanta-based property management firm with about 55 employees reportedly saved around $100,000 annually by building a custom replacement using Claude Code and Replit. Against that backdrop, Slack CMO Ryan Gavin sat down for an exclusive interview with VentureBeat to frame the announcement and argue that the company's future depends on an idea he calls "multiplayer AI" — and that the 25 years of customer data locked inside Salesforce is an asset no vibe-coded alternative can replicate. Why Slack's CMO believes 'multiplayer AI' is the next big enterprise battleground Gavin's core argument is that the enterprise AI conversation has been stuck in single-player mode for too long, and that Slack is uniquely positioned to break it open. "So much of what we've seen are just these incredible tools that have largely been single-player, incredible tools for individual productivity, helping people complete tasks and write code," Gavin told VentureBeat. "But as we've always known at Slack ever since our inception, work is a team sport. For AI to really take hold in the enterprise, it has to be multiplayer." The distinction matters commercially. Most AI assistants today — ChatGPT, Claude, Copilot — default to one-on-one conversations with a single user. A researcher queries a model, gets a response, and acts on it alone. The insight stays in a private chat window, invisible to colleagues. Gavin argues this creates a new version of the tab-switching problem that plagued pre-AI enterprise software, except now employees are also navigating dozens of individual agent interfaces on top of their existing applications. "It's going to benefit almost no one if every enterprise application out there spawns hundreds of agent babies, and employees end up in a worse world than they were before," Gavin said. Slack's answer is to make Slackbot the orchestration layer. Because everything happens in shared channels, any action an agent takes — pulling a customer profile, flagging a deal risk, updating a Jira ticket — is visible to the entire team. A colleague can redirect, build on, or correct the agent's work in real time. How MCP and Salesforce's headless 360 platform power Slackbot's new capabilities The technical backbone of the announcement is the Model Context Protocol , an open standard originally developed by Anthropic that defines how AI models discover and invoke external tools. MCP has seen rapid adoption across the AI tooling ecosystem. By early 2026, it had been adopted by Claude Code , Cursor , GitHub Copilot , and OpenAI's tooling, with managed hosting available from AWS , Cloudflare , and Vercel . As a DEV Community explainer puts it, MCP "is the closest thing the AI tooling ecosystem has to a standard." In this implementation, Salesforce exposes its platform capabilities — CRM records, Tableau visualizations, Data 360 customer profiles, Agentforce agents — as MCP servers. Slackbot operates as an MCP client, connecting to those servers and routing user queries to the appropriate back-end system. When a user asks Slackbot about a customer, the bot discovers which MCP tools are relevant, calls them, and synthesizes the results into a single response — all within the Slack conversation. Gavin explained the architecture in simple terms: "Salesforce is extending what has always been our open platform through our Headless 360 strategy — making all of these MCP endpoints available. And then Slackbot acts as an MCP client, connecting to those MCP servers and bringing all that data in within the confines of a trusted permission platform." That permission layer is critical. Slackbot respects each user's Salesforce permissions, meaning a marketing coordinator cannot accidentally access sales pipeline data they are not authorized to see. Validation rules, field-level security, and org-wide data boundary configurations carry over automatically. For admins, setup requires no custom integration code — Salesforce MCP servers can be discovered, installed, and governed from a single UI using the existing Slack-Salesforce connection. Salesforce first introduced the Headless 360 concept at its TDX developer conference in April, positioning it as an API-driven layer that exposes the platform's data, workflows, and governance controls so that software agents, rather than human users, can execute business processes directly. As CIO.com reported at the time, analysts viewed the move as an effort by Salesforce "to position itself as a central layer for managing agent-driven operations across different business functions." Slack says it's betting on openness, not on any single AI protocol When asked whether Slack is making a risky bet on MCP as a protocol — given that standards in AI tooling can shift rapidly — Gavin reframed the question entirely. "We're not betting on MCP, per se. We're betting on what we've always bet on, which is that Slack is an open platform," Gavin told VentureBeat. "MCP happens to be the best agent-to-agent protocol that the industry is rallying around right now, but if something better came out tomorrow, you'd see the same pattern from Slack — we're going to stay open. MCP and APIs are simply tools that facilitate that." That open-platform philosophy is central to Slack's identity and, Gavin argues, its competitive differentiation. Slack already hosts more than 2,600 app integrations . The new MCP-native partner ecosystem includes Atlassian , Box , DocuSign , Canva , Lucid , Zoom , and more than 25 additional companies, each of whose agents can be added directly to shared Slack channels. MuleSoft Agent , now connected to Slackbot, helps manage integrations for the team — checking system health or surfacing critical error alerts in the same workspace where the team is already collaborating. But MCP is not without trade-offs. The protocol requires tool discovery on every connection, and large tool libraries can consume significant context tokens. One technical analysis noted that a server exposing 300 tools could cost 5,000 to 10,000 tokens per session before the model does any useful work. For an enterprise like Salesforce with hundreds of potential tools across CRM, analytics, and service platforms, careful filtering and segmentation of MCP servers become essential design decisions — a challenge the company will need to navigate as the ecosystem scales. Inside Slack's complicated relationship with Anthropic and the Claude question Perhaps the most delicate topic in the interview concerned Slack's relationship with Anthropic, the AI lab behind Claude — and one of Slack's most visible power users. Just last week, Anthropic launched Claude Tag , a persistent AI teammate that works inside Slack channels, prompting confusion among Salesforce employees who worried it competes directly with Slackbot and Agentforce. The Information reported internal anxiety about whether Salesforce was welcoming a competitor into its own living room. Salesforce has financial reasons to maintain the partnership: the company reportedly expects to spend $300 million on Anthropic tokens this year and holds a stake in Anthropic. Gavin addressed the tension head-on, framing it as a feature of Slack's platform strategy rather than a threat. "We're incredibly excited and bullish about what Anthropic is bringing into Slack. Period. End of statement," Gavin said. He noted that Anthropic "is building roughly 65% of their code with Claude in Slack," and pointed out that ChatGPT was originally built in Slack, as was Perplexity. "Building nowadays happens in the open, and every company is going to be building in the open with tools like this, and you need a platform to build in the open," Gavin said. His argument is that feature overlap between Slackbot , Claude Tag , and other third-party agents is "actually a feature, not a bug" — a sign of a healthy platform rather than a competitive vulnerability. He compared it to an ecosystem where multiple products serve similar needs but win on craftsmanship, ease of use, and integration depth. "One of the reasons Slackbot has been the fastest-adopted feature in Salesforce history is the simplicity, the approachability — underpinned by the trust that comes from having an agent that knows me, knows my tone, knows my work, knows my people, knows my data," Gavin said. The distinction Slack draws is structural: Slackbot has access to a user's full workspace context, Salesforce data, permissions, and connected applications by default. Claude Tag, by contrast, only sees the channels it is explicitly added to. For Slack's leadership, that asymmetry is the moat. How Slack plans to compete with Microsoft Teams and Google in the AI era Asked directly about competitive positioning against Microsoft Teams and Google Workspace , Gavin pointed to Slack's open channel architecture as the differentiator no competitor can replicate. "If you spend any time in Teams, it's a lovely tool for chat, direct messages, and video, but it has no platform for open communication across organizations," Gavin said. "Its SharePoint-based architecture is fundamentally limiting." He cited Shopify as an example, where an internal AI agent called River is deployed across approximately 4,400 channels serving 6,000 employees. He also referenced a Fortune report noting that Microsoft's own head of AI mandated that his team run on Slack rather than Teams — a pointed detail Gavin clearly relished. "There's a reason for that," he said. "We're in an era right now where openness matters, and all the other tools you mentioned, they're still relatively closed." The competitive pressure is real and intensifying. Microsoft has integrated Copilot across its entire productivity suite, giving it a distribution advantage that reaches virtually every Fortune 500 company. Google has been similarly aggressive with Gemini across Workspace. And new entrants are crowding the market: a startup called Viktor , which embeds AI agents inside Slack and Teams workspaces, recently raised a $75 million Series A led by Accel — with Slack cofounders Stewart Butterfield and Cal Henderson participating as angel investors. Box , one of the enterprise customers highlighted in the announcement, told Slack it aims to have its sellers complete 75 to 80 percent of their work inside Slack. Gavin repeated that figure as evidence that the platform is becoming the default workspace for entire organizations, not just engineering teams — a shift he believes accelerates as AI makes every employee a builder. Slack's biggest long-term play is making Salesforce's CRM useful to everyone in the company Gavin saved what he considers the most underappreciated element of the announcement for last: the democratization of Salesforce's CRM. For 25 years, Salesforce's CRM has been used primarily by sales, service, and marketing professionals — a relatively modest percentage of a company's total workforce. The promise of Slackbot as a conversational interface is that any employee, regardless of their role or technical fluency, can now query and act on CRM data simply by asking a question in natural language. "What most people don't realize is that this democratization of CRM is going to take its usage from a modest percentage of employees to the entire enterprise," Gavin said. "When you can make systems like Data 360 or Agentforce for Sales accessible to the entire employee base — not just a percentage — think about how much more valuable those investments become." He cited Engine , a company that handles 800,000 customer inquiries a year, as an example. Previously, answering a customer inquiry required a specific employee with access to a specific tool to look up a customer's history. Now, anyone in the company can ask Slackbot and see a complete customer profile, review case history, and write updates — all without being retrained or learning a new interface. Engine's CEO Elia Wallen, in a statement sent to VentureBeat, described the integration as enabling employees to "make data-driven decisions and take action without leaving the conversation." The financial logic is straightforward: if Salesforce can make its platform useful to 100 percent of a customer's workforce rather than the 20 or 30 percent who currently hold licenses, the value of the existing Salesforce investment multiplies without requiring a proportional increase in spending. That pitch becomes especially potent at a time when CIOs are scrutinizing every line of their AI budgets. What analysts and CIOs should watch as Slack rolls out its biggest AI update yet The announcement is a significant architectural evolution for Slack, but several questions remain unanswered. First, pricing. The company did not directly address whether Slackbot's MCP-powered Salesforce integration will require additional SKUs or license tiers. As Info-Tech Research Group analyst Scott Bickley cautioned when Headless 360 was first announced in April, "Salesforce's MO seems to be to announce new capabilities that require SKUs. CIOs should be asking about pricing now." Second, performance. Routing user queries through MCP servers to Salesforce back-end systems introduces latency that could affect the conversational feel Slack prides itself on. Neither the press release nor the interview disclosed SLAs for MCP tool calls — a gap that enterprise buyers will want addressed. Third, the competitive dynamics of the platform play. Slack's open-platform philosophy invites powerful partners like Anthropic and OpenAI into its ecosystem, but those same partners are building their own surfaces for enterprise work. Anthropic reportedly plans to expand Claude Tag to Microsoft Teams, email, and other project management tools — meaning the partner Salesforce is paying hundreds of millions a year is building the infrastructure to be useful without Slack at all. And fourth, the broader existential question facing all enterprise software: whether AI agents will ultimately reduce the need for CRM systems entirely. Gavin's pitch — that Slack makes CRM more valuable by making it more accessible — is the inverse of the bear case. The market will ultimately decide which thesis prevails. Salesforce reported record first-quarter revenue of $11.1 billion in fiscal Q1 2027 , with Agentforce ARR surpassing $1 billion for the first time and combined AI and data ARR reaching $3.4 billion. Those numbers suggest the AI strategy is beginning to generate real revenue, even as the company navigates a market that remains uncertain about the long-term trajectory of legacy enterprise software. "Slack has quickly moved from this beloved collaboration tool from the last ten years to now this multiplayer AI platform that we call a work operating system," Gavin said. Five years ago, Salesforce paid $27.7 billion for what was, at its core, a very good group chat application. On Wednesday, it started trying to prove that group chat was never the product — it was the foundation. In the age of AI agents, the most valuable real estate in enterprise software may not be the database where the data lives. It may be the conversation where the decisions get made.
- Toyota-backed Chinese self-driving unicorn Momenta hits Hong Kong market
Toyota-backed Chinese self-driving unicorn Momenta hits Hong Kong market Nikkei Asia
- Anthropic's fix for Fable 5's high cost is turning it into a manager that delegates to Sonnet 5
Anthropic recommends using the expensive Claude Fable 5 mainly as a planner for smaller models instead of running it on every task. Combined with Sonnet 5 in the "Advisor" pattern, this setup hits 92 percent of Fable 5's solo performance at 63 percent of the cost. The article Anthropic's fix for Fable 5's high cost is turning it into a manager that delegates to Sonnet 5 appeared first on The Decoder .
- SpaceXAI wants to compete on AI infrastructure, not just AI models
SpaceX is best known for its outer space and rocket projects, while xAI has largely been focused on its flagship AI assistant, Grok. These are two very different paths, but now the two are one. This week, Elon Musk announced SpaceXAI , which brings xAI and SpaceX together as a company. The re-branding seems to indicate that the company plans to push even harder into AI and the critical infrastructure that underpins it. But experts say enterprise buyers should remain cautious. “SpaceXAI is becoming a credible player in AI infrastructure, but it is not yet at the stage where most enterprises should consider it a primary AI provider,” said Jehaan Nanavaty , a senior advisory analyst at Info-Tech Research Group. He added that established AI providers like Microsoft and OpenAI, AWS and Anthropic, and Google “continue to lead” in areas like governance, regulatory compliance , enterprise support, and ecosystem maturity. Space is the ‘only way to scale’ Elon Musk founded xAI in March 2023. It was acquired by SpaceX in February 2026, ultimately unifying the billionaire’s AI and space ambitions. SpaceX said at the time its intention was to “form the most ambitious, vertically-integrated innovation engine on (and off) Earth,” consisting of AI, rockets, space-based internet, and direct-to-mobile device communications. In addition to building out Grok’s capabilities, xAI has continued to develop Colossus , which it said is the world’s largest and most powerful AI supercomputer. Located in Memphis, Tennessee, it is built on an interconnected cluster of roughly 200,000 Nvidia H100 GPUs, constructed, the company said, in just 122 days. SpaceX acquired xAI to overcome the “immense” power and cooling constraints of AI data centers here on earth, the company said. Musk argued that global electricity demand for AI “simply cannot be met with terrestrial solutions, even in the near term.” “In the long term, space-based AI is obviously the only way to scale” and resource-intensive efforts should be shifted to locations with vast possibilities, the company said, noting “space is called ‘space’ for a reason.” The newly-minted SpaceXAI combines rocket and satellite manufacturing and AI infrastructure, and the company plans to build data centers in space powered by solar energy. It says it will deploy “AI compute satellites” as early as 2028. It is rumored to be releasing its first jointly-developed AI model with its recent acquisition, Cursor, this week. SpaceX’s IPO filing revealed that it spent $12.7 billion on AI in 2025, more than 3x its investment in its other business units. Around the same time as that IPO, the company filed “ Boosting America’s Space Economy ” with the US Federal Communications Commission (FCC), which detailed its plan to build a constellation of up to one million satellites that would operate as orbital data centers and rely on solar power to run their onboard computing systems. Along the way, SpaceX has inked some notable AI infrastructure deals: Anthropic has agreed to pay $1.25 billion per month for access to Colossus, while Google has signed a deal worth $920 million a month. SpaceXAI codifies AI ambitions Info-Tech’s Nanavaty pointed to SpaceXAI’s strategy of combining Grok models, Colossus GPU clusters, Starlink networks, and SpaceX launch capabilities as something that provides a “level of vertical integration that can’t be easily replicated.” “If SpaceXAI executes on its roadmap, it could emerge as a serious competitor by differentiating on infrastructure rather than model performance alone,” he said. The company’s most distinguishing quality is its intersection of AI and space infrastructure, Nanavaty noted. It is the “clear leader in mass-to-orbit launch capacity, with no real competition,” and further, Starlink has already demonstrated its ability to manufacture, deploy, and operate satellites at an “unprecedented scale.” “If any organization is capable of building orbital AI infrastructure, it is SpaceXAI,” he said, adding that its $55 billion investment in the 11-million-square-foot Gigasat factory will further strengthen that position. That build is set to begin as soon as late 2027. In the long term, space-based AI compute could enjoy benefits like abundant solar power, reduced dependence on terrestrial energy infrastructure, and the ability to process data directly in orbit, Nanavaty noted. That said, the concept remains “largely unproven,” and significant engineering challenges, particularly around servicing and maintaining hardware in space, still need to be addressed. Further, while SpaceX has a “strong track record” of delivering ambitious engineering projects, its timelines have often slipped, sometimes by several years, said Nanavaty. “Demo systems by 2028 appear realistic,” he noted, but large-scale commercial deployments are likely to take longer. This is because both the technology and the business case will need to mature before orbital AI data centers become a viable alternative to terrestrial infrastructure. Thus, he advised, “be cautious about assigning a firm timeline beyond early demonstrations.” This article originally appeared on NetworkWorld .
Score: 67🌐 MovesJul 8, 2026https://www.cio.com/article/4194200/spacexai-wants-to-compete-on-ai-infrastructure-not-just-ai-models-2.html - What a harness is and how to build one with Claude Agent SDK
Listen now | 🎙️ I built a custom Claude Agent SDK harness to automate Sentry bug triage, saved the "dear agent, please fix this" prompt forever, and show you exactly how to build your own
- OpenAI launches GPT-Live, a full-duplex voice upgrade that lets ChatGPT talk more like a person
OpenAI on Wednesday launched GPT-Live , a pair of new voice models that fundamentally redesign how people talk to ChatGPT — replacing the company's existing Advanced Voice Mode with an architecture that can listen and speak simultaneously, much like an actual human conversation. The two models, GPT-Live-1 and GPT-Live-1 mini , are rolling out globally starting today across iOS, Android, and ChatGPT.com. GPT-Live-1 becomes the default voice model for paid ChatGPT users on the Go, Plus, and Pro tiers, while GPT-Live-1 mini serves free-tier users. OpenAI also plans to bring the models to the API, and developers can sign up to be notified. The release marks the third generation of ChatGPT's voice technology in roughly two years — and OpenAI's clearest bid yet to turn its chatbot into something that feels less like querying a search engine and more like talking to a colleague. Why full-duplex voice changes everything about talking to AI The defining technical advance in GPT-Live is what OpenAI calls a " full-duplex architecture ." In telecommunications, full-duplex means both parties on a phone call can talk and listen at the same time. Applied to AI, it means the model continuously processes your incoming audio even while it generates its own spoken response — no more waiting for a clean silence gap to figure out when you've finished a thought. "Instead of processing a sequence of separate messages, GPT-Live continuously processes input while generating output," OpenAI wrote in its research blog. "The model can therefore make interaction decisions many times per second: whether to speak, continue listening, pause, interrupt, or invoke a tool." In practice, that translates to a voice assistant that can insert conversational acknowledgments — "mhmm," "yeah," "got it" — while you're still talking, pick up on a natural pause without jumping in prematurely, and handle rapid interruptions without derailing the entire exchange. OpenAI's previous Advanced Voice Mode , launched to paid users in September 2024, processed and generated audio within a single model but still operated on rigid turn-by-turn exchanges. As OpenAI acknowledged in the announcement, "because turn detection is based on silence, even a brief pause or background noise could be mistaken for the end of turn — causing the model to interrupt at unnatural times." That brittleness created a product that, while impressive in demos, could be deeply frustrating in extended real-world use. Background chatter in a coffee shop could trigger a response. A thinking pause might get swallowed. The experience felt, as one researcher put it on X shortly after the announcement, like " walkie-talkie turn taking ." GPT-Live is designed to end that era. How OpenAI split voice and intelligence into two separate layers GPT-Live introduces a second structural change that may prove just as consequential for enterprise adoption: it decouples the voice interaction layer from the reasoning layer. When a user asks a straightforward question, GPT-Live handles it directly. But when the query demands web search, deeper reasoning, or more complex agentic work, GPT-Live delegates the task to a frontier model running in the background — at launch, GPT-5.5, the large language model OpenAI released in April — and continues talking with the user while the computation happens asynchronously. "While it works, GPT-Live can keep talking with you and maintain the flow of conversation," OpenAI explains. "As we release new frontier models, we'll continuously update the model used by GPT-Live." This delegation model is a meaningful architectural bet. Rather than building a single monolithic voice model that tries to be both conversationally fluid and deeply intelligent, OpenAI has split the problem in two: a voice-native model optimized for real-time interaction, and a separate reasoning engine that can be swapped out as the state of the art improves. It is, in effect, a modular design — one that allows OpenAI to upgrade the intelligence of its voice assistant without retraining the voice model itself. The implications for enterprise and developer workflows are significant. A voice agent built on this architecture could maintain a natural conversation with a customer while simultaneously querying databases, searching the web, or performing multi-step reasoning — tasks that would have introduced several seconds of dead air under the old pipeline. The three generations of ChatGPT voice, from clunky pipeline to continuous stream To understand how far voice AI has come, it helps to trace the three generations that led to GPT-Live . The original ChatGPT Voice , launched in 2023, used a cascaded pipeline — a speech-to-text model ( Whisper ) transcribed what you said, a large language model ( GPT-4 ) generated a text response, and a text-to-speech model converted that response back into audio. Each handoff introduced latency and lost information. As OpenAI noted, "the complexity came at a cost: information could be lost across models, and responses were slow and stilted." That cascaded approach was the industry standard, and its limitations were well-documented. As the blog OpenHelm noted in an October 2024 analysis of OpenAI's Realtime API, the old pipeline stacked up to roughly 1,700 milliseconds of latency — nearly two full seconds of dead air before the first word of a response. Managing the state between the three separate APIs consumed an enormous amount of engineering effort. OpenAI's Advanced Voice Mode, which began its limited rollout to paid ChatGPT Plus users in July 2024 before expanding more broadly in September 2024, collapsed that three-model pipeline into a single model that processed audio natively. As TechCrunch reported at the time, the rollout came with five new voices — Arbor, Maple, Sol, Spruce, and Vale — alongside improved accent handling and smoother conversations. The feature also launched on the web in November 2024, extending it beyond mobile. But Advanced Voice Mode still operated through discrete, alternating turns — and it launched into the shadow of a PR debacle that OpenAI is still working to leave behind. The Scarlett Johansson controversy still shadows OpenAI's voice ambitions Advanced Voice Mode arrived in the wake of one of OpenAI's most damaging self-inflicted crises. During the GPT-4o launch in May 2024, the company showcased a voice called "Sky" that many listeners immediately noted sounded strikingly similar to Scarlett Johansson , who famously voiced an AI companion in the 2013 film Her . Johansson said she had declined OpenAI CEO Sam Altman's offer to voice the system, then was "shocked, angered and in disbelief" when the product launched with a voice her own friends couldn't distinguish from hers, as NBC News reported. Altman had tweeted just the word "her" the day the product launched. OpenAI pulled the voice and apologized, but the incident drew public scrutiny from SAG-AFTRA and members of Congress , and crystallized broader concerns about AI companies moving fast with creative IP. The Hollywood labor union said the issue underscored "why we're strongly championing federal legislation that would protect their voices and likenesses ... from unauthorized digital replication," as NBC News reported . Forbes contributor Paul Tassi wrote at the time that Altman, "by holding up Her on a pedestal of something to strive for, has missed the point of that film" — in which the protagonist's relationship with his AI companion ultimately does him more harm than good. GPT-Live appears designed, in part, to move past those controversies. OpenAI says it has "remastered the nine distinct voices in ChatGPT for GPT-Live" and notes the system "is designed for conversation, not voice impersonation," with "safeguards to prevent it from imitating a real person's voice." What 150 million weekly voice users will actually notice today OpenAI disclosed that more than 150 million people talk to ChatGPT using voice and dictation features each week — a notable slice of the platform's 900 million total weekly active users. The voice experience has grown into a substantial product in its own right, used for language practice, bedtime stories, commute-time chat, and hands-free everyday help. The new product features reflect that usage. GPT-Live introduces rich visual cards that surface during voice conversations — weather forecasts, stock data, sports scores, and maps — giving users something to glance at without breaking the flow of speech. Users can now choose between three reasoning levels for answers: Instant for quick responses, Medium for moderate thinking, and High for more complex work. And if you take a moment to think, "ChatGPT Voice now waits instead of jumping in and interrupting," OpenAI wrote. "If you ask it to stay quiet and listen, it will. And when there's background noise, like passing traffic or nearby conversations, ChatGPT is better at focusing on your voice instead of getting distracted." Early reactions from users with preview access were cautiously positive. "I had early access to sol. it is a phenomenal model," wrote one user on X , adding it is “much better at frontend, long context knowledge work, and its vibes are much better.” Another observer cut to the heart of the matter: "The smarts are not new here, GPT-Live hands hard questions to GPT-5.5. What is new is the feel: full-duplex voice that listens while it talks." New voice-specific safety tests reveal where the risks still live The GPT-Live system card , published alongside the announcement, reveals a safety strategy built around the particular risks of real-time voice interaction — a domain where the speed and intimacy of conversation create hazards that text-based chat does not. OpenAI expanded its safety evaluations to include audio-native tests, using both real user voice samples (from those who opted in) and synthetically generated prompts targeting edge cases across categories like self-harm, sexual content, illicit behavior, emotional reliance, mental health, and hate speech. On the synthetic evaluations — which OpenAI described as deliberately adversarial — GPT-Live-1 showed substantial improvements over Advanced Voice Mode. In illicit behavior, for instance, the safety score rose from 0.63 to 0.97. On self-harm, it climbed from 0.72 to 0.98. Hate speech achieved a perfect 1.00, up from 0.87. On the production-prompt evaluations — which used real user audio and reflected more ambiguous, borderline scenarios — the picture was more mixed. GPT-Live-1 matched or improved on Advanced Voice Mode in most categories but showed a slight regression on emotional reliance (from 0.88 to 0.82), though OpenAI noted the change was not statistically significant. The company built real-time safeguards that can intervene while the model is speaking — steering toward safer responses, surfacing crisis resources, or ending the voice conversation entirely in higher-risk situations. It also designed additional protections for teen users and adapted self-harm support flows for voice, including crisis helpline integration. Perhaps most notably, OpenAI said it is "rolling out longer-term measurement and post-launch monitoring focused on emotional reliance" — an acknowledgment that the very naturalness GPT-Live strives for creates its own category of risk. Google, ByteDance, and Nvidia are already in the full-duplex race While OpenAI was refining its safety guardrails, its rivals were shipping full-duplex systems of their own. Google's Gemini Live , which supports full-duplex conversation alongside camera and screen sharing — capabilities GPT-Live notably lacks at launch — is already available in the Gemini app. Google released Gemini 3.1 Flash Live in March as its highest-quality real-time audio model, targeting low-latency voice interactions for developers. ByteDance launched Seeduplex in April, claiming to be the first production-scale full-duplex speech AI deployed at scale, inside its Doubao app. Seeduplex reported roughly a 50 percent reduction in false-response and false-interruption rates compared to ByteDance's previous half-duplex system. And Nvidia's PersonaPlex , released in January, brought customizable voice and role control to full-duplex models, breaking what had been a constraint where natural-sounding models were locked into a single fixed voice. The competitive picture is clear: full-duplex voice interaction is quickly becoming table stakes for consumer AI products, not a differentiator. OpenAI's advantage lies in the scale of its existing user base, its integration with GPT-5.5's reasoning capabilities, and the breadth of the ChatGPT ecosystem. But the window in which any one company has a monopoly on natural-sounding voice AI has already closed. OpenAI also acknowledged several gaps. GPT-Live does not support voice with video or screen sharing at launch. Language support is limited, with the company noting that "for certain languages, the model may have a non-native accent or gaps in fluency." And API access is not available on day one, meaning enterprise developers cannot yet build on GPT-Live directly — a constraint that will slow the model's penetration into commercial voice-agent workflows where competitors like Google, ElevenLabs, and Deepgram already have developer-facing products. The end of the chat box may be closer than anyone expected GPT-Live is essentially OpenAI's most significant bet yet on voice as the primary interface for AI — not just a convenience feature bolted onto a text chatbot, but a purpose-built interaction layer that sits between the user and the company's most powerful models. "Over time, we believe this research will also unlock the ability to use voice for increasingly complex, longer-running, and more agentic work," OpenAI wrote. That ambition — using natural voice as the front end for autonomous AI agents that can perform multi-step tasks — is the logical endpoint of the full-duplex plus delegation architecture. Imagine telling your phone to book a flight, negotiate with your insurance company, or debug a production server, all through a conversation that feels as natural as talking to an assistant who also happens to have the intelligence of a frontier AI model. Two years ago, talking to ChatGPT meant dictating into a microphone and waiting nearly two seconds for a stilted reply. One year ago, it meant a smoother exchange that still felt like a polite, slightly awkward phone call with someone who insisted on waiting for you to finish every sentence. Today, it means something closer to a real conversation — imperfect, still constrained in some languages and missing video, but unmistakably closer. OpenAI once got into trouble for wanting to recreate the movie Her . With GPT-Live, the company may finally be reckoning with the harder question the film actually posed: not whether AI can sound human enough to talk to, but what happens to us when it does.
- Physical AI ‘space race’: can Europe compete with China and the US in humanoid robotics?
European firms say they are fighting to secure a foothold in physical AI – the integration of artificial intelligence into robotics and machinery – as China and the United States take an early lead in the sector, with industry insiders warning the continent faces the threat of further deindustrialisation if it fails to establish a competitive industry. “You see China and the US … because of AI … typically they are considered the leaders, but do not count out Europe,” said David Kehr, president...
- AI-backed cyberattacks are exposing the limits of enterprise VPNs: Report
As AI helps attackers accelerate cyberattacks and attackers hand off compromised access in seconds, traditional VPN-based defences are now facing stern challenges
- Suspecting AI cheating, Ivy League prof ordered an in-person final; scores fell 50%
AI cheating leads to "a failed society," professor says.
- Polysense raises $10.7M to scale AI quality control for food manufacturers
Belgian startup Polysense has raised $10.7million in an oversubscribed seed funding round to accelerate the globalrollout of its AI-powered quality control platform for food manufacturers. Theround wa...
Score: 65💰 MoneyJul 8, 2026https://tech.eu/2026/07/08/polysense-raises-107m-to-scale-ai-quality-control-for-food-manufacturers/ - Adobe Advertising Just Launched Its Own Custom Algorithms Product
Adobe paid $540 million to buy video DSP TubeMogul in 2016 with plans to build a programmatic advertising powerhouse inside one of the world’s largest marketing clouds. But that promise has gone largely unfulfilled during the 10 years since. Until now, perhaps. Last week, Adobe Advertising announced the general release of its own Custom Algorithms […] The post Adobe Advertising Just Launched Its Own Custom Algorithms Product appeared first on AdExchanger .
Score: 65🌐 MovesJul 8, 2026https://www.adexchanger.com/marketers/adobe-advertising-just-launched-its-own-custom-algorithms-product/ - This Robotaxi Company Is Growing Its Share Of US Market — Not Waymo, Not Tesla
Depending on which side of the fence you sit on, either Waymo or Tesla is leading the robotaxi revolution in the United States. Objectively, Waymo has far more self-driving cars and robotaxi users, but Tesla fans think only Tesla has the right approach long term, and that improvements in “Full ... [continued] The post This Robotaxi Company Is Growing Its Share Of US Market — Not Waymo, Not Tesla appeared first on CleanTechnica .
- A Beginner’s Guide to Amazon Bedrock: Your First LLM App Without the Overwhelm
From zero AWS knowledge to a working Claude-powered app in one sitting. Continue reading on Towards AI »
- Research: AI Is Changing What Employers Want from New Hires
Three capabilities increasingly distinguish high-potential talent.
- AI's biggest challenge is not compute - it's data storage
SPONSORED FEATURE: As AI evolves from novelty to autonomy, the real bottleneck isn't processing power—it's where to put all that data.
Score: 65🌐 MovesJul 8, 2026https://www.theregister.com/storage/2026/07/08/ais-biggest-challenge-is-not-compute-its-data-storage/5267453 - ‘GitLost’ Flaw Lets Attackers Trick GitHub AI Agent Into Leaking Private Repos
‘GitLost’ Flaw Lets Attackers Trick GitHub AI Agent Into Leaking Private Repos DevOps.com
Score: 65🌐 MovesJul 8, 2026https://devops.com/gitlost-flaw-lets-attackers-trick-github-ai-agent-into-leaking-private-repos/ - ActiveCampaign Launches Google Ads Connector for Active Intelligence, Bringing AI-Guided Campaign Creation and Reporting to Marketers
Marketers and business owners can now create and manage Google Ads Performance Max campaigns without leaving ActiveCampaign — part of a growing ecosystem of Active Intelligence connectors that brings the full power of a marketer’s tech stack into one AI-powered workspace
- How to build a voice agent that transfers to a human
Step-by-step guide on creating a voice agent that can hand off calls to a human operator.
- Gutenberg launches AI Motion Studio for faster, scalable video production in digital and CTV era
Gutenberg has launched AI Motion Studio for digital and CTV video production. This new engine compresses production timelines from six weeks to just one day. It converts a single narrative into over ten platform-ready video formats. The studio combines AI capabilities with human creative direction and expertise. This innovation aims to close the creative production gap in the evolving media landscape.
- ‘기술 관리자’는 끝났다…AI 시대 CIO가 갖춰야 할 6가지 리더십
AI는 조직의 모든 계층에서 업무 수행 방식은 물론, 업무를 담당하는 주체까지 바꾸고 있다. 이에 따라 경영진의 역할과 리더십 방식도 달라지고 있다. 이제 경영진은 AI를 활용해 조직을 새롭게 설계하고, 그 과정에서 수반되는 불확실성을 관리해야 하는 과제를 안고 있다. 이러한 변화는 CIO의 역할에도 영향을 미치고 있다. CIO는 새로운 책임을 맡고 더 높은 기대를 받으면서 역할이 확대되고 있다. 이는 수년간 이어져 온 CIO 역할의 진화 과정의 연장선에 있다. 과거 기술 운영을 책임지는 관리자였던 CIO는 전략적 비즈니스 혁신을 지원하는 역할을 거쳐, 이제는 조직의 미래 비전을 제시하는 리더로 자리매김하고 있다. 베테랑 CIO와 연구자, 업계 자문가들은 오늘날 IT 리더에게 요구되는 새로운 리더십 원칙 6가지와, 이를 대체하게 된 기존 리더십 원칙을 소개했다. CEO 보고에서 비전 공동 수립으로 기존 원칙: CEO에게 보고한다. 새로운 원칙: CEO와 함께 조직의 미래 비전을 만든다. 기업 경영의 오랜 관행은 CEO가 조직의 궁극적인 목표와 방향을 제시하면 CIO를 비롯한 다른 경영진이 이를 실행하기 위한 계획을 수립하는 방식이었다. 프로티비티(Protiviti)의 CIO 솔루션 총괄 매니징 디렉터 샤론 스터플비미 (Sharon Stufflebeme)는 “이제 CIO는 CEO와 긴밀히 협력해 조직의 미래 비전을 함께 만들어야 한다”라고 설명했다. 이어 “미래를 내다보는 통찰력과 함께 그 변화가 현재 조직에 어떤 영향을 미칠지 이해해야 한다. 또한 현재의 조직을 미래에 맞게 전환할 방안을 마련하고, 앞으로 현실적으로 일어날 변화를 예측해 조직이 어떻게 적응할지 구체적인 방향을 제시할 수 있어야 한다”라고 말했다. 스터플비미는 “이러한 역량은 과거에도 중요했지만 CIO에게 가장 핵심적인 역량은 아니었다”라며 “하지만 이제 CIO는 AI를 비롯한 새로운 기술이 창출할 가치와 그에 따른 비용, 위험을 가장 잘 이해하는 위치에 있다. 따라서 조직의 미래 비전을 수립하고 이를 실현할 실행 방안을 제시하는 역할까지 맡아야 한다”라고 밝혔다. 성과 지원에서 미래 설계자로 기존 원칙: 비즈니스 성과를 지원한다. 새로운 원칙: 미래의 조직을 설계한다. 지난 몇 년 동안 최고경영진은 CIO에게 AI를 이해하기 쉽게 설명하고, 이를 활용해 비즈니스 성과를 창출하는 방안을 제시해 줄 것을 기대해왔다. 그러나 MIT 슬론 CIO 심포지엄(MIT Sloan CIO Symposium) 집행의장 앨런 테이트 (Allan Tate)는 이제 경영진의 기대 수준이 한 단계 더 높아졌다고 진단했다. 이제는 CIO가 AI를 활용해 조직 자체를 새롭게 설계하기를 기대한다는 것이다. 테이트는 “이제 질문은 ‘AI로 무엇을 할 수 있는가’가 아니다. ‘AI를 활용해 조직을 어떻게 다시 설계할 것인가’가 핵심”이라며 “‘AI를 책임감 있고 효과적으로 활용하는 조직을 어떻게 만들 것인가’가 CIO에게 주어진 새로운 과제다. CIO는 점차 조직 혁신을 설계하는 아키텍트 역할을 맡게 될 것”이라고 말했다. 그는 이러한 역할을 수행하려면 CIO가 불확실성과 긴장을 관리하며 조직을 이끌 수 있어야 한다고 설명했다. 테이트는 “CIO는 불확실성을 편안하게 받아들일 수 있어야 한다”라며 “올바른 질문을 던지고, 다양한 관점에서 문제를 해석하며, 서로 다른 이해관계와 긴장을 조율하고, 인간의 지능과 AI를 어떻게 조화롭게 결합할지 고민해야 한다”라고 말했다. 이어 “동시에 조직 구성원도 불확실성에 익숙해질 수 있도록 도와야 한다. 모든 사람이 같은 의견에 도달하는 상황은 기대하기 어렵다. CIO에게 필요한 것은 불확실한 환경에서도 더 나은 경영 판단을 내리는 능력이다. 또한 구성원이 위협을 느끼지 않고 모두가 함께 성장할 수 있다고 믿을 수 있는 신뢰의 문화를 구축해야 한다”라고 설명했다. 테이트는 AI가 업무를 자동화하면서 일자리가 줄어들 수 있다는 우려를 인정하면서도, CIO는 경영진과 함께 AI 기반 혁신이 만들어낼 새로운 역할과 새로운 기회를 고민해야 한다고 강조했다. 그는 “모든 기술 혁명에서 그랬듯 가장 어려운 일은 앞으로 어떤 새로운 일이 생겨날지를 상상하는 것”이라고 말했다. 실패 장려에서 심리적 안전으로 기존 원칙: 빨리 실패하라(Fail Fast). 새로운 원칙: 구성원이 안심하고 성장할 수 있는 환경을 조성하라. ‘실패를 두려워하지 말고 빨리 실패하라’는 것은 기술 업계에서 가장 많이 회자된 원칙 중 하나였지만, 실제로는 제대로 구현되지 못한 경우가 많았다. 이제 IT 리더에게 요구되는 것은 가능성이 낮은 프로젝트를 서둘러 포기하는 데 그치지 않는다. 구성원이 실패를 안전하게 받아들이고, 그 과정에서 얻은 교훈을 축적해 더 빠른 혁신으로 이어갈 수 있는 환경을 만드는 것이 중요해졌다. 분석 장비 및 소프트웨어 기업 워터스(Waters)의 수석부사장 겸 CIO 브룩 콜란젤로 (Brook Colangelo)는 글로벌 IT 조직을 운영하면서 ‘간단한 진단법’을 활용하고 있다고 소개했다. 콜란젤로는 “팀의 성과가 기대에 미치지 못하거나 변화에 저항하는 상황이 발생하면 구성원의 다섯 가지 심리적 욕구 가운데 무엇이 위협받고 있는지를 먼저 살펴본다”라며 “그 다섯 가지는 지위(Status), 확실성(Certainty), 자율성(Autonomy), 관계성(Relatedness), 공정성(Fairness)이며, 이를 직접적이고 공감하는 방식으로 해결하려고 노력한다”라고 설명했다. 이러한 접근은 조직 문화에 기반을 두고 있다. 콜란젤로는 “워터스 IT 조직은 동기 부여와 성장에 관한 신경과학적 원리를 바탕으로 운영된다”라며 “성과는 함께 축하하고, 실패는 원인을 분석하며, 그 경험을 팀 전체의 학습 기회로 활용한다”라고 말했다. 그는 이러한 심리적 위협을 진단하고 해결하는 능력이 오늘날 CIO에게 가장 중요한 리더십 역량 가운데 하나라고 평가했다. 특히 “IT 조직은 본질적으로 구성원이 위협을 느끼기 쉬운 환경이며, AI 시대에는 이러한 특성이 더욱 두드러진다”라고 설명했다. 콜란젤로는 “이러한 역량을 갖추기까지는 시간이 걸렸지만 의도적인 교육과 훈련을 통해 조직에 정착시킬 수 있었다”라며 “IT 리더십 포럼(IT Leadership Forum)을 통해 조직의 리더들이 이러한 행동을 직접 실천하고 이를 조직 전체로 확산할 수 있도록 지원했다”라고 말했다. 그는 팀 문화에 대한 이러한 투자가 IT 조직이 네 가지 대규모 전략 과제를 동시에 추진할 수 있었던 원동력이 됐다고 평가했다. 해당 과제는 ▲인수 기업 통합 ▲인도 글로벌 역량 센터(Global Capability Center) 직원의 정규직 전환(수락률 99%) ▲ERP를 SAP S/4HANA로 전면 전환 ▲조직 전반의 AI 혁신을 안전하게 추진하기 위한 기반 구축이다. 콜란젤로는 “각 프로젝트는 사람마다 서로 다른 반응을 불러일으킨다”라며 “심리적 위협 신호를 공통된 언어로 이해하면 무엇이 조직의 속도를 늦추고 있는지 정확히 진단하고, 그 문제를 직접 해결할 수 있다”라고 말했다. 비즈니스를 아는 CIO로 기존 원칙: 비즈니스 전문가와 협업한다. 새로운 원칙: 비즈니스 전문가가 된다. CIO는 오래전부터 기술만 알아서는 성공할 수 없다는 사실을 깨달았다. 이에 각 사업 부문의 담당자와 협력하며 비즈니스의 문제점과 현안을 파악하고, 다른 경영진과 함께 사업 부문별 목표와 전략을 이해하는 데 힘써왔다. 하지만 이제 CIO는 한 단계 더 도약해야 한다고 리더십 자문 및 임원 채용 전문 기업 위트키퍼(WittKieffer)의 IT·디지털 리더십 부문 총괄 파트너 제프 스터먼 (Jeff Sturman)은 말했다. CIO가 최고운영책임자(COO)처럼 조직 운영 전반을 이해하는 역할로 진화해야 한다는 것이다. 스터먼은 “이제 CIO는 전략, 운영, 고객 경험 등 조직의 모든 활동이 만나는 중심에 서 있다”라며 “비즈니스의 모든 영역과 연결되는 역할”이라고 설명했다. 이어 “CIO는 여전히 기술과 보안, 그리고 AI 분야에서 최고의 전문성을 갖춰야 한다. 하지만 이제는 COO처럼 조직 운영 전반을 이해해야 한다. 오늘날 IT 리더의 손길이 닿지 않는 비즈니스 영역은 사실상 없기 때문”이라고 말했다. 예를 들어 의료 분야 CIO라면 사업 운영뿐 아니라 규제 요건, 임상 운영 등 다양한 영역까지 폭넓게 이해해야 한다고 그는 설명했다. 물론 다른 경영진도 비즈니스를 잘 알아야 한다. 그러나 AI 도입을 IT 조직이 주도하면서 업무 자동화와 업무 혁신이 빠르게 진행되는 만큼 CIO는 다른 경영진보다 조직 전체의 운영 방식과 업무 프로세스를 더욱 깊이 이해해야 한다고 스터먼은 강조했다. 그는 아직 모든 CIO가 이러한 수준에 도달한 것은 아니지만, 점점 더 많은 IT 리더가 조직 운영 전반을 한눈에 조망하는 이른바 ‘파노라마식 시각(panoramic view)’을 갖추고 있다고 평가했다. 재무를 아는 CIO에서 CFO형 CIO로 기존 원칙: 조직의 재무를 충분히 이해한다. 새로운 원칙: CFO처럼 사고한다. 레드햇의 수석부사장 겸 CIO 마르코 빌 (Marco Bill)은 많은 CIO와 마찬가지로 이전보다 훨씬 더 많은 재무 분석 업무를 수행하고 있다. 클라우드와 AI 투자 비용을 효율적으로 관리하면서도 성능 저하 없이 비용을 절감할 수 있는 방안을 찾기 위해서다. 예를 들어 빌과 그의 팀은 각 워크로드를 퍼블릭 클라우드에서 운영하는 것이 유리한지, 프라이빗 클라우드가 적합한지, 아니면 자체 데이터센터에서 운영하는 것이 비용 효율적인지를 지속적으로 분석하고 있다. 일부 워크로드를 온프레미스 환경으로 이전해 2,000만 달러 이상(약 301억 원)의 비용을 절감했으며, 이러한 의사결정은 모두 재무 분석을 기반으로 이뤄졌다고 설명했다. 빌은 “이러한 계산은 한 번으로 끝나는 것이 아니라 지속적으로 반복해야 하는 작업”이라고 말했다. 스터플비미 역시 AI 프로젝트가 확대되면서 CIO가 재무 분야에 더욱 깊이 관여하게 될 것으로 내다봤다. CEO와 이사회가 AI 투자에 대해 정량적으로 입증 가능한 투자수익률(ROI)을 요구하고 있기 때문이다. 그는 “IT는 조직이 나아갈 비전을 제시하는 것은 물론 어떤 투자가 ROI를 창출할 수 있는지 보여줄 수 있는 재무적 역량도 갖춰야 한다”라며 “이제 CIO는 어디에서 가치가 발생하고 어떤 비용이 수반되는지를 정확히 이해해야 한다. 이러한 역량은 원래도 중요했지만 AI 시대에는 그 중요성이 더욱 커졌다”라고 말했다. 스터플비미는 지금까지 AI 투자에서 충분한 ROI를 확보하기 어려웠고, 실패한 AI 프로젝트에 대한 경영진의 인내심도 갈수록 줄어들고 있다고 지적했다. 그는 “이사회와 CEO는 가치가 어떻게 창출되는지 이해하고, 이를 정량적으로 측정하며, 실제로 그 가치를 실현할 수 있는 CIO를 원한다”라고 설명했다. 또한 AI 에이전트가 일부 사람의 업무를 대신하게 되면 비용 구조 역시 달라질 것이라는 점을 이해해야 한다고 강조했다. 스터플비미는 “에이전트가 비용 자체를 없애는 것은 아니지만 비용 구조는 바꾼다”라며 “따라서 CIO는 이러한 새로운 역량의 총소유비용(TCO)을 산정할 수 있어야 한다. 이는 자사뿐 아니라 협력사에도 적용된다. CIO는 협력사로부터 얻는 가치가 지불하는 비용보다 큰지를 판단할 수 있어야 한다”라고 말했다. 리더가 아닌 팀원에 맞춰라 기존 원칙: 직원이 자신의 리더십 스타일에 맞추기를 기대한다. 새로운 원칙: 팀원에게 맞춰 리더십 스타일을 바꾼다. 전략 기술 컨설팅 기업 태펫 어소시에이츠(Taffet Associates)의 매니징 파트너 겸 CIO 그레그 태펫 (Greg Taffet)은 이제는 자신의 리더십 스타일과 조직 구성원과 소통하는 방식을 팀원에게 맞게 조정해야 한다고 말했다. 태펫은 “전 세계 곳곳에 팀원이 있는 만큼 예전처럼 사무실에서 자연스럽게 얼굴을 마주하며 일하던 시절과는 관리 방식이 크게 달라졌다”라고 설명했다. 그는 리더로서 구성원이 어떤 방식과 환경에서 가장 생산적으로 일할 수 있는지를 이해하려고 노력한다고 말했다. 완전 원격 근무를 선호하는 사람도 있고, 서로 다른 시간대에 비동기적으로 일하는 방식을 원하는 사람도 있으며, 정해진 일정에 맞춰 사무실에서 근무하거나 원격과 출근을 병행하는 하이브리드 근무를 선호하는 사람도 있기 때문이다. 태펫은 “사람마다 최고의 성과를 내기 위한 조건은 모두 다르다”라며 “모든 사람이 재택근무에서 높은 생산성을 내는 것도 아니고, 반대로 모두가 항상 사무실 근무에서 가장 높은 성과를 내는 것도 아니다”라고 말했다. 또한 그는 문화적 배경이나 개인적 특성이 서로 다른 리더십 방식에 어떤 영향을 미치는지 이해하고, 각자의 강점을 최대한 끌어낼 수 있는 환경을 만드는 데도 힘쓰고 있다고 설명했다. 태펫은 “학교에서 학생의 학습 방식이 시각형인지, 청각형인지, 체험형인지에 따라 수업 방식을 달리하듯, 이제 리더십도 구성원 개개인에게 맞춰야 한다”라고 말했다. dl-ciokorea@foundryco.com
- The enterprise AI copilot playbook for business leaders
Deploy enterprise AI copilots with this playbook covering search, agents, governance, ROI, and department use cases for business leaders.
Score: 65🌐 MovesJul 8, 2026https://www.glean.com/blog/the-enterprise-ai-copilot-playbook-for-business-leaders - AI Will Redefine B2B Marketing Models — Get Ahead Of The Shift At AI Forum Singapore
Buyer behavior is changing faster than most go-to-market models can adapt. At AI Forum Singapore, learn how to align strategy, execution, and AI adoption to compete in a world of autonomous buyers.
- ByteDance, Tencent, and Alibaba Are Shutting Down AI Chat Companions — Here's Why
China three biggest tech firms are dismantling the AI companion features that millions of users built emotional attachments to, citing regulatory and strategic reasons.
Score: 65🌐 MovesJul 8, 2026https://pandaily.com/bytedance-tencent-alibaba-ai-companion-shutdown-jul2026 - Datacentres are a ticking timebomb. We must make sure AI’s benefits outweigh the costs | Nicki Hutley
They suck up energy and water, and blast out heat. Just who is better off from all this investment – aside from tech bros? The two great existential threats of our time – the climate crisis and AI – come hurtling together in the explosion of datacentres across Australia and around the world. You can hardly avoid hearing about them these days, either with awed reverence of the promised benefits to humankind or with fear and anger given the implications for the climate, inflation, jobs and even housing affordability. Continue reading...
- AI has collapsed the cyber response window — resilience now starts before the attack
Presented by Rubrik Enterprise cybersecurity is facing a fundamental speed problem. Frontier AI models are now enabling autonomous attacks that can move from initial access to full system breakout in as little as 27 seconds . That’s faster than any human-operated security workflow can detect, escalate, and respond. As a result, security operations can no longer assume there is time for humans to respond between breach and damage. The security posture that enterprises need for the AI era centers on cyber resilience: continuously identifying clean recovery states, mapping critical data and identity dependencies, and automating restoration so that operations can recover in hours not days. "Everything that relied on process or human-in-the-loop intervention is no longer going to be able to execute at the speed of the attacks," says Dev Rishi, GM of AI at Rubrik. "If the attacks are happening in 27 seconds, it means I need my recovery to happen just as quickly." Traditional detection and prevention are failing against AI-driven attacks The rules-based logic that has defined enterprise security for decades, such as static access controls, known signature detection and deterministic behavioral policies, was engineered for deterministic software. AI agents behave differently. They're non-deterministic, capable of pursuing the same objective through many different paths, and increasingly capable of circumventing static guardrails by finding alternative routes when one is blocked. The deeper problem is that conventional security logic checks identity, permissions, and access, and asks whether each individual access is permitted. But it can’t evaluate whether a sequence of permitted actions, taken across multiple applications, constitutes either a data leak, a destructive operation, or an attack. "You need a system that can understand context," Rishi says. "You need to use AI to look at what an agent is doing and say, ‘it looks like what you're doing might be a risk of leaking sensitive data externally.’" How AI agents are blurring the line between internal and external cyber threats Enterprise security has historically maintained a meaningful distinction between external and internal threat vectors. External threats can be multidimensional, lightning fast, and come from a variety of vectors. On the other hand, internal threats were traditionally bounded by what a single human actor could accomplish before detection, constrained in speed, scope, and scale, but that distinction is falling apart as AI agents operate inside enterprise environments. These agents have access to multiple systems simultaneously and move at speeds no human employee can match. When an agent makes a mistake, such as a hallucination, misread instruction, or an unintended data transfer, the resulting damage can look operationally identical to a malicious insider attack. And when an external attacker compromises an internal agent, they inherit its full access profile across every connected application. "Whether or not the agent is an internal threat because of an inadvertent mistake or because it's been maliciously compromised, you need runtime guardrails that enforce your organizations policies consistently across agents," Rishi says. "The practical answer is an AI-native guardian layer that monitors agent behavior semantically, understands intent across actions, and can block or terminate a misbehaving agent at machine speed, then trigger recovery immediately." Preparing for a world of inevitable compromise Frontier AI models, including those capable of discovering and operationalizing zero-day vulnerabilities autonomously, are changing the economics of attacks. As a result, interest in Mythos readiness is growing. Enterprises are increasingly operating under two assumptions: that attacks are inevitable, not exceptional, and that investment in resilience and rapid recovery must be treated as strategically as investment in prevention has been. The shift reframes recovery from a post-incident activity into a capability that is deliberately designed, tested, and continuously validated. "The idea that you can recover quickly from an attack is going to become one of the most important facets of security," Rishi says. "It's the insurance policy that organizations now have to treat as a first-class citizen." Why AI-powered cyber resilience depends on small models True cyber resilience is a two-sided coin: it demands both real-time intelligent enforcement to intercept threats in motion, and automated recovery to restore operations immediately. While having backups is a baseline, organizations need workflows that can continuously monitor systems at machine speed, and instantly determine the most recent clean state under attack conditions. Applying AI to the first half of that equation—real-time enforcement—creates a fundamental technical and economic challenge. Relying on massive frontier models to monitor every agent action introduces crippling latency overhead and exorbitant computing costs. A guardian AI system that slows down operations or costs as much as the systems it monitors is simply not viable for widespread adoption. “It has to be a fast, small, and cheap AI model,” Rishi says. “No one wants to sign up for a secure solution that doubles their cost or latency.” This is why small language models (SLMs) are critical for real-time enforcement. Rubrik’s approach, anchored by its acquisition of Predibase, is to build this frontline defense layer on small models optimized specifically for speed and efficiency. Unlike heavy frontier models, SLMs can semantically evaluate agent behavior at machine speed and at a fraction of the cost, acting as a real-time checkpoint. That hyper-efficient enforcement layer is what enables a tighter, seamless connection to recovery. When the system observes an agent taking a destructive action—such as deleting a database, corrupting a critical file, or exfiltrating sensitive data—the small model detects it immediately, halts the damage, identifies the most recent clean snapshot from before the incident, and initiates recovery in a single, automated workflow. The shift from incident response to architectural resilience The broader implication of Mythos and similar frontier AI systems is a shift in how organizations think about security. As AI compresses the gap between attack and impact, resilience and recovery become architectural requirements rather than operational considerations. Rubrik’s view is that security systems can no longer stop at detection. As AI agents gain greater autonomy, observability, identity context, and recovery must operate as a coordinated resilience layer. The goal is not simply to identify when something has gone wrong, but to shorten the gap between detection and restoration. "The same thing that's introducing the threats, the frontier capabilities of models like Mythos, can also be used to help us combat the threat," Rishi says. "Positioning yourself for the AI era means closing the gap between detecting that something has gone wrong and restoring the systems that were affected, before the cost of that gap compounds." Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. For more information, contact sales@venturebeat.com .
- Tesla’s Robotaxi Push Tests New Blueprint for Scaling Fast
Tesla’s Robotaxi Push Tests New Blueprint for Scaling Fast The Information
Score: 65🌐 MovesJul 8, 2026https://www.theinformation.com/articles/teslas-robotaxi-push-tests-new-blueprint-scaling-fast - Global AI industry falls short on safety, think tank warns
Global AI industry falls short on safety, think tank warns The Straits Times
- Artificial intelligence (AI) is Advancing Faster Than Governments Can Protect People, New Global Index Finds
Artificial intelligence (AI) is Advancing Faster Than Governments Can Protect People, New Global Index Finds
- Claude Cowork expands to mobile and web as Anthropic reveals what people actually use it for
Anthropic makes Claude Cowork available on web and mobile, with cloud option now available, as usage patterns shift.
- Can AI equalize political campaign ads – or will it remain a tool for spreading lies?
Political campaigns are increasingly deploying AI and deepfakes to further their messaging, and the scale of spread has experts concerned From the comfort of his bed, Jonathan Rinaldi, a political candidate for a city council seat in Queens, New York, tinkered away on his iPhone, prompting an artificial intelligence chatbot to mock up fake news hits and endorsements he had never received. During the campaign last October, Rinaldi shared one of those stories, made to appear real with a CNN logo, on his Facebook and Instagram. It stated that Lynn Schulman, his opponent and an incumbent Democrat, had been “forced to drop out of the race due to a series of critical mistakes”. But Schulman had not quit her campaign, and in November, won by a landslide. Continue reading...
Score: 63🌐 MovesJul 8, 2026https://www.theguardian.com/technology/2026/jul/08/ai-ads-political-campaigns - Temasek targets big jump in AI investments as value of portfolio hits record high
Temasek targets big jump in AI investments as value of portfolio hits record high Reuters
- IBM and Red Hat Expand Lightwell with New Offerings to Build the Trust Infrastructure for AI-Era Open Source
RALEIGH, N.C. and ARMONK, N.Y. — IBM and Red Hat today announced the commercial launch of Lightwell, delivering automated vulnerability remediation at scale through two offerings: Lightwell Network and Lightwell Clearinghouse Premier. Available now, Lightwell Network gives enterprises access to a launch catalog of 6,500+ remediated, digitally signed, and certified application-layer dependencies across major ecosystems,... … continue reading The post IBM and Red Hat Expand Lightwell with New Offerings to Build the Trust Infrastructure for AI-Era Open Source appeared first on SD Times .
- The UN has an AI strategy for everyone except the labs that build it
The Geneva Digital Week opened July 6 with the inaugural Global Dialogue on AI Governance and the AI for Good Global Summit, where the United Nations’ new Independent International Scientific Panel on Artificial Intelligence presented to governments its first global scientific assessment of AI . The gathering caps three years in which the U.N. has produced an impressive volume of work on AI, from the Global Digital Compact to the Governing AI for Humanity report ; from UNESCO’s recommendation on the ethics of AI to the International Telecommunication Union’s annual summits . Read together, this work shares a single posture in which the U.N. treats AI as something to be received, a downstream resource to be channeled toward beneficial ends, aligned with the Sustainable Development Goals, monitored for societal effects, fitted with ethical guardrails. This is the demand side of technology, and it’s where all the U.N.’s substantive engagement currently sits. The supply side, or the places where frontier AI is produced, evaluated, and released, has no meaningful U.N. presence at all. There is no multilateral body with technical staff who can examine a laboratory’s work, no arrangement for evaluating training runs, no shared infrastructure for incident reporting across borders. The governing architecture for the next several decades is consolidating right now in bilateral arrangements between frontier labs and the governments that host them, in private entities like Anthropic’s Project Glasswing and in export-control decisions by the parties hosting the technology. Once those institutional facts are established, the path of least resistance for every subsequent decision will be to extend them rather than to build a multilateral alternative. The pattern is visible in the news cycle. The U.S. Commerce Department recently authorized the release of Anthropic’s most-able AI model to roughly 100 American institutions, two weeks after an export-control suspension had taken it offline for everyone. Both the suspension and the selective release were decided by the White House. The approved partners are American, but the populations affected by the technology are global. European officials have publicly expressed frustration at this new dependence on decisions made in Washington, D.C. The authority to determine which populations receive access to a frontier-safety technology now resides, by default, in a single national administration. What would a supply-side role for the U.N. look like? Pre-deployment evaluation of AI models is happening. Britain’s AI Security Institute and the U.S. Center for AI Standards and Innovation (formerly the U.S. AI Safety Institute), both test frontier models through voluntary agreements with the major laboratories, a practice so routine that the labs themselves cite it as evidence of responsible development. The question that the multilateral system has so far failed to ask is why the safety assurance available to two countries should not be available to the other 191 states across the world. The U.N. has done this level of technical work before. The International Atomic Energy Agency (IAEA) was created in 1957 to engage directly with the places where nuclear material is made, and its safeguards give it a standing technical presence in countries that have signed the Nuclear Non-Proliferation Treaty. The agency works inside the facilities where nuclear material is handled, rather than from outside the industry that handles it, and the populations of countries without nuclear programs receive assurance from a system that none of them could build on their own. The AI frontier labs have reasons of their own to engage. They are already submitting models to voluntary evaluation, because it produces a credential they can cite and a defense against the argument that frontier development is happening without oversight. A multilateral arrangement extends the same logic, giving the labs a guarantee they can offer to markets and governments beyond the two countries that currently provide it and protecting them from the regulatory fragmentation that country-by-country deals would otherwise produce. They have an interest in being trusted by the world rather than only by Washington and London. The difficulty is that frontier AI production is concentrated in the U.S. and China, and neither shows much appetite for opening its labs to a multilateral presence. The same was true of nuclear safeguards in 1953, four years before the IAEA was operational, when a similar concentration of capability in two rival powers, then Washington and Moscow, appeared to make any arrangement politically impossible. The agency was built because both sides eventually concluded that mutual visibility was preferable to mutual opacity, and the short window between 1953 and 1957 turned out to be the formative period during which the architecture was set for the next 70 years. The UN80 reform agenda is the obvious vehicle through which supply-side oversight could emerge, since the initiative is grappling with how to refocus the U.N. Secretariat for future multilateralism, and the AI governance gap is exactly the structural absence a reform process is meant to address. The U.N. cannot exercise authority on behalf of the billions of people it serves if it is present only on the side where the consequences of frontier technology are felt and absent from the side where those consequences are produced. Filling that gap does not require a new agency on day one: A UN80 mandate for a small standing-evaluation unit, assembled from assets the U.N. already owns, such as the International Computing Center , and an invitation to the labs to extend to the center the voluntary agreements they already honor in London and Washington would be enough to begin. That is the work the next phase of multilateral engagement will have to take up. A version of this essay originally appeared on PassBlue .
- Voice AI Platform: A Business Guide to Smarter Customer Conversations
My friend runs a mid-size insurance company. Good product, solid team, genuinely cares about its customers. But about two years ago, he told me something that stuck with me. He said, “ We’re losing people not because we did anything wrong, but because they just got tired of waiting. ” Not waiting for a resolution. Waiting for someone to pick up. That single sentence sums up the problem better than any industry report I’ve read. Customers aren’t holding your company to some impossibly high standard. They’re just comparing you to the last app they used, the last delivery they tracked, the last thing that worked without friction. And a phone call that opens with four minutes of hold music and a cheery automated menu? That’s not going to win. This is where the conversation around a voice AI platform starts making practical sense, not as a tech trend, but as a real fix for a real headache that businesses have been complaining about for decades. So let’s talk about what it actually is, how it’s used, what separates a good rollout from a bad one, and what your business specifically needs to consider before spending a dollar on it. What is a Voice AI Platform, Really? Strip away the marketing language, and you get something pretty straightforward. A voice AI platform is a system that picks up the phone and handles the conversation start to finish, or at least far enough that a human only gets involved when it genuinely matters. It listens, understands, responds, and does things like pulling account info, updating records, or booking appointments, all in real time without a person sitting in a chair somewhere. Now, this is different from those phone trees that have been annoying people since the 1990s. Those systems are basically flowcharts; they only work if you follow the path they expect. Say something slightly off-script and the whole thing collapses. Voice AI actually processes what a person is saying and figures out the intent behind it, not just the keywords. The building blocks that make it work: ASR: Automatic Speech Recognition: Turns spoken words into text as the caller speaks. Accuracy has gotten remarkably good, even with accents, background noise, or fast talkers. NLU: Natural Language Understanding: The piece that figures out what someone actually wants. “ I need to move my appointment ,” and “ can we reschedule Thursday’s thing? ” mean the same thing. NLU connects those dots. TTS: Text-to-Speech: How the AI responds out loud. The robotic voices of five years ago are basically gone. Modern TTS is genuinely hard to distinguish from a person in casual conversation. Dialog Management: Keeps the conversation coherent across multiple turns so the AI doesn’t forget what was said two exchanges ago. The Learning Layer: Every call teaches the system something. Over months, a well-maintained platform becomes meaningfully better at handling your specific customers and your specific types of calls. None of that requires you to have an AI team in-house. Most platforms are built so that people without technical backgrounds can configure, review, and improve them. Why This Topic Is Getting So Much Attention Right Now There’s a reason the conversation around voice AI shifted from “ interesting idea ” to “ serious operational priority ” over the past few years. A few things converged at once. Contact center hiring became a genuine crisis. Turnover rates in the industry hover somewhere around 30–45% annually in many markets. That means companies are perpetually training new people, perpetually dealing with service quality dips, and perpetually burning money on a problem that doesn’t seem to get better, no matter how much attention it gets. Meanwhile, call volumes didn’t drop; they went up. More products, more channels, more customer expectations across the board. Companies are being asked to do more with workforces that keep shrinking. And on the other side of the phone, customer tolerance for a bad experience has compressed dramatically. People don’t write angry letters anymore. They leave a one-star review, post something on social media, and switch providers sometimes all three before they’ve even gotten off the call. The drivers pushing businesses toward a Voice AI Platform are pretty consistent across industries: Per-call costs with live agents typically run between $6 and $15, depending on complexity. At any real volume, that adds up to millions annually, and automation can cut that number by 60% or more for routine calls. Coverage gaps at nights and weekends, when call centers operate with skeleton crews, and customers calling during those windows often get the worst experience. The “ boring call ” problem , where trained, relatively expensive agents spend most of their day answering the same five questions over and over, which tanks morale and wastes their actual skills. A generational shift in customer preference, a significant chunk of callers, particularly under 40, genuinely prefer resolving simple things themselves without talking to a person. They don’t want empathy for a balanced inquiry. They want it done in 45 seconds. These aren’t futuristic problems. They’re today’s problems for most companies with any real inbound call volume. The Features That Separate Good Platforms from the Ones That Get Abandoned If you’ve ever sat through a voice AI vendor demo, you’ve probably noticed they all look great under controlled conditions. The real question is what happens when your actual customers call in tired, distracted, with weird questions, using slang, switching topics mid-sentence. Here’s what to actually probe when you’re comparing options: Does It Handle Messy, Real Conversations? Demos are polished. Customers aren’t. The best way to stress-test a platform before you sign anything is to throw real scenarios at it, pull five or ten actual transcripts from your call logs, the confusing ones, and run them through whatever the vendor is showing you. How does it handle a caller who gives half an answer, pauses, then changes what they were asking? What’s the recovery when the AI misunderstands? Does it spiral, or does it gracefully ask for clarification? Accents, fast speech, background noise: Does it stay coherent, or does accuracy fall off a cliff? How Deep Does the Integration Actually Go? “Integrates with your CRM ” is probably the phrase I’d ban from vendor conversations if I could. It says almost nothing. Ask them to show you live, not in slides, what the integration does during a call. Can it pull up a customer’s account, current order, outstanding balance, or service history in real time while the conversation is happening? Does it write outcomes back to the system when the call ends, or does someone still have to do that manually? If your CRM is something custom or less common than Salesforce, what does that actually look like? Escalation Design: The Most Underrated Piece Here’s a strong opinion: a voice AI system that can’t get out of someone’s way when they need a human is worse than no AI at all. Nothing makes a customer angrier than feeling trapped in a loop with a machine that keeps misunderstanding them and won’t let them leave. What triggers an escalation: specific words, failed attempts, detected tone, or caller request? When it escalates, what does the agent receive? A full transcript? A summary? Or does the customer have to start over from scratch? Can a caller bail out to a human at any point, without having to fight for it? Reporting That Actually Helps You Improve Most platforms will hand you a dashboard with dozens of metrics. Maybe two or three of them are things you’d actually act on. Know what you’re looking for before you go shopping: Resolution rate by call type, not just overall averages. Where in the conversation are calls failing or escalating? That’s where the tuning happens. Sentiment trends over time. If customers are consistently getting frustrated at a specific point, that’s a signal worth knowing about. Transcript search so you can pull actual examples, not just statistics. Compliance is Non-Negotiable if You’re in a Regulated Industry If your business touches health data, financial accounts, or personal information in any meaningful way, this needs to be the first conversation, not an afterthought. HIPAA certification matters for healthcare. Ask for documentation. PCI-DSS applies anytime payment card data crosses the line. Verbal assurances aren’t enough. Find out exactly where recordings are stored, how long they’re kept, and who can access them. Voice biometrics for identity verification. Some platforms have it, and for certain industries, it’s both more secure and more convenient than the usual security questions. A Realistic Deployment Playbook (Without the Vendor Spin) The technology side of a voice AI rollout is usually the part that goes fine. The part that goes sideways is everything else: picking the wrong starting point, skipping conversation design, rushing to full deployment, then wondering why satisfaction scores dropped. Here’s what works: Start With the Calls Nobody Wants to Take Open your call records for the last quarter. What are the five most common reasons people are calling? There’s a 90% chance that at least half of them are questions with straightforward, consistent answers: order status, appointment scheduling, balance checks, store hours, and basic account changes. These are your runway. Not the complex stuff. Not the emotional calls. The predictable, repeatable ones that your best agents could handle in their sleep but probably wish they didn’t have to. Sit With Your Actual Agents Before Writing a Single Line of Dialogue This step gets skipped constantly, and it’s the reason so many deployments produce an AI that sounds like it was written by a committee of people who’ve never worked a phone shift. Your agents know things that aren’t in any documentation. They know the weird way customers phrase a specific question. They know what phrase always means someone is about to ask three more questions. They know where conversations stall and why. Spend an afternoon with them before you build anything; you’ll save yourself months of retraining later. Stress-Test It Before Any Customer Gets Near It Get people in the building or outside it to try to break the system. Different accents, different phrasing, bad audio, weird edge cases, and intentionally confusing inputs. Document every failure, fix the obvious ones, and decide which edge cases you’re going to escalate vs. try to handle. This isn’t just quality assurance. It’s how you find out where the experience is going to frustrate people before it happens in the real world. Roll It Out Gradually, Not All at Once Start by routing maybe 20% of the relevant call type through the AI. Keep the rest on live agents. Run both in parallel for a few weeks, compare every metric you can, and use that data to tune. Then expand. This approach feels slower. It is slower. It also means you don’t have a mass customer service failure on your hands because something broke at full scale in week one. Make Someone Responsible for It Every Week A voice AI platform isn’t software you install and ignore. It needs a person who doesn’t have to be a developer, could be a supervisor or a smart ops manager who is reading transcripts regularly, catching where the system is missing the mark, and feeding that back in. The platforms that quietly become incredible over time all have this person. The ones that plateau or quietly get abandoned don’t. Where Real Businesses Are Getting Real Results Theory is fine. Specifics are better. Retail and E-commerce Order tracking is the bread-and-butter use case here. “ Where’s my package? ” calls are completely automatable. The AI looks up the order, reads the status, gives an estimated delivery, and handles the occasional exception without drama. One mid-size retailer I read a case study on redirected 40% of their inbound call volume away from live agents just by nailing this one use case. Not 40% of all calls, just the order-related ones. That’s still a significant number. Order and delivery status, including exception handling Return initiation and status updates Store hours, locations, and product availability by SKU or location Healthcare The no-show problem in healthcare is genuinely expensive, and reminder calls with built-in rescheduling capability address it directly. A patient who gets a call the day before realizes they can’t make it and reschedules in that same conversation; that’s a slot that would have otherwise gone empty. Simple idea, measurable result. Appointment reminders with live rescheduling built in Pre-visit intake and insurance confirmation Post-discharge follow-up check-ins Prescription pickup and refill notifications Financial Services Security is the first conversation here, and it should be. But modern voice AI platforms built for financial use cases have the compliance architecture to handle it. Voice biometrics, for what it’s worth, are actually more secure for identity verification than asking someone’s first pet’s name and faster. Balance inquiries, transaction history, and recent charges Fraud alert confirmation and card status Loan application and disbursement status Scheduled payment reminders and processing Hospitality Front desk call volume during check-in hours is brutal. A lot of them are questions that don’t need a trained staff member to answer, such as parking instructions, breakfast hours, early check-in availability, and nearby restaurant recommendations. Automating those frees up the actual hospitality staff to do the work that matters. Reservation confirmation, changes, and cancellations Pre-arrival instructions and logistics Concierge-type information at scale Post-stay feedback collection Telecom Telecom customers are often already frustrated by the time they call. Outages, billing confusion, service issues- these aren’t happy-caller situations. Voice AI handles the informational and transactional calls well, which means when a human does get involved, they’re not fielding call number 300 about the same outage. They’re actually solving something. Network outage updates and status checks Bill inquiry and payment processing Basic device troubleshooting flows Upgrade eligibility and plan comparisons What Happens to Your Team When You Bring This In I want to address something directly because it comes up in almost every conversation about automation: no, this isn’t about replacing your people. That might sound like the obvious thing to say, and I understand the skepticism. But here’s the practical reality: the call types that voice AI handles well are the ones that burn agents out. Nobody goes into customer service because they are passionate about reading order statuses. The calls that actually require a human, the upset customer who needs someone to really listen, the complex billing dispute, the situation that doesn’t fit any standard script, those still land with people. When businesses deploy Voice AI for Customer Support , the consistent feedback from actual agents isn’t anxiety; it’s relief. Relief that the 47th identical question of the day isn’t coming to them. Relief that when a call does arrive, it comes with a full summary of everything that already happened, so they’re not starting from zero. Agent-side changes worth knowing: Pre-call context arrives automatically, the agent sees who’s calling and what was discussed before they say a word. Post-call documentation generates itself, which gives back 5 to 10 minutes per call that used to go to manual note-taking. Burnout indicators drop when agents aren’t fielding repetitive volume all day. Training actually improves because AI transcripts surface gaps in knowledge base coverage that random call monitoring never would have caught. The best voice AI implementations treat agents and automation as a relay team, not competitors for the same work. Choosing a Vendor Without Getting Burned The market is noisy. Startups with great demos, enterprise players with long sales cycles, everything in between. Here’s how to actually narrow it down. Push Back on the Demo Don’t let them script it entirely. Bring your own messy, real-world scenarios and ask them to run through those instead. A platform that performs beautifully under vendor-designed conditions but stumbles on your actual calls isn’t production-ready for you. What’s your accuracy rate on natural, unscripted speech, real number, not a benchmark? What languages and accents are your training data actually built on? How long from signing to live calls, specifically? Ask for a project timeline, not an estimate. Who owns the conversation data? Can we delete it? Can you use it to train other clients’ models? Describe your support process when something breaks at 11 pm on a Saturday. Warning Signs Worth Taking Seriously Can’t or won’t do an unscripted live demo. Basic integrations require significant custom development. Pricing that looks fine at low volume, but the math doesn’t hold at scale. Vague or evasive answers about data jurisdiction and ownership. No case studies from companies in your industry or your size range ask for references, not logos. How Pricing Tends to Work Three models dominate the market, and your choice should depend on how predictable your call volume is: Per-minute billing is straightforward at lower volumes, but expensive when volume spikes, and hard to budget for. Subscription-based by concurrent call capacity, with predictable monthly costs, better for operations with consistent volume. Hybrid a base platform access fee with consumption pricing on top, which spreads risk but requires careful modeling before you commit. The Numbers Conversation At some point, these lands end up on someone’s desk for budget approval, and the question becomes: what does this actually do for the business financially? Here’s a realistic comparison based on what mature deployments actually report: Metrics: The harder-to-quantify side of the ledger matters too. Customers who get resolved faster churn less. Outbound proactive calls, reminders, renewal notices, and delivery updates happen automatically instead of falling through the cracks. Agents who aren’t exhausted perform better on the calls they do take. A voice AI platform at full deployment isn’t just a cost line; it’s an operational asset. The Mistakes That Kill These Projects There are predictable failure modes in voice AI deployment, and most of them aren’t technical. Going too wide, too fast. Trying to automate 60% of call volume immediately, before the system is trained on your specific customers and call types, is how you generate a customer service disaster and a cancelled contract. Underestimating the importance of tone. An AI that resolves the issue but sounds cold, robotic, or bizarrely chipper will still leave customers feeling like they had a bad experience. Voice, pacing, and personality matter even in automated calls. Skipping transcript review. The transcript review loop is where a platform gets good. Without someone reading them and identifying where things go wrong, the system stays at whatever level it launched at, which is never the final goal. Designing the escalation path last. Escalation should be designed first, not bolted on at the end. How a call moves from AI to human, and what the human receives when it does, is as important as anything else in the system. Treating launch as completion. The day a voice AI platform goes live is the day it starts accumulating data you can use to make it better. Businesses that treat deployment as a finish line miss almost all of the value. Conclusion When my friend with the insurance company finally switched over to an AI-assisted call system, he told me the first thing he noticed wasn’t the cost savings, though those came. It was that his team stopped dreading Mondays. The calls that had been grinding people down just stopped showing up in the live queue. What remained were the ones that actually required a person. That shift from contact center as a burden to contact center as a place where the interesting work happens is something a lot of businesses aren’t expecting when they go into a voice AI platform deployment. They’re expecting efficiency metrics. They get those too. But the morale piece tends to surprise people. The businesses doing this well aren’t the ones who bought the most expensive platform. They’re the ones who started with a specific problem, designed carefully around it, and kept their hands on the wheel after launch. There’s no shortcut to that. But there’s also nothing exotic about it. It’s the same discipline that makes any operational improvement work. If you’re evaluating this for your business, the practical advice is to stop comparing vendors on paper and start testing them on your actual call scenarios. The difference between what sounds good in a deck and what works with your real customers is where the decision actually lives. Voice AI Platform: A Business Guide to Smarter Customer Conversations was originally published in Towards AI on Medium, where people are continuing the conversation by highlighting and responding to this story.
- Women and university graduates in Australia most at risk of losing jobs to AI, report finds
Those with high levels of vocational training, including tradespeople, are least exposed to AI displacement, according to government review Get our breaking news email , free app or daily news podcast Artificial intelligence has yet to cause widespread job losses but the federal government has warned that telemarketers, advertising staff and accountants are among the occupations “most exposed” to being replaced by the technology. According to a first-of-its-kind national report , people in the more exposed occupations are more likely to be women and have university qualifications. Continue reading...
- Anduril’s billionaire founder warns U.S. colleges are falling behind China—where students are learning AI
Anduril’s billionaire founder warns U.S. colleges are falling behind China—where students are learning AI Fortune
- Bug in top AI coding agents shows that Unix-era security headaches never really die
'GhostApproval' problem highlights human-in-the-loop fails
- China's Momenta debuts flat in Hong Kong as IPO caution grips AI listings
China's Momenta debuts flat in Hong Kong as IPO caution grips AI listings Reuters
- A self-driving Waymo just reported its own passengers to police
If you think no one is paying attention inside your driverless taxi, think again. Two California teenagers were detained Monday after a Waymo detected suspicious behavior during their ride and alerted police in San Mateo. The teenagers were allegedly spotted drinking alcohol and using toy guns to shoot Orbeez—small, squishy jelly-bean-like water beads—out the windows, leading Waymo to park in a nearby lot and lock the passengers inside until authorities showed up. A video posted to social media shows around five armed officers and a K9 unit circling the Waymo vehicle, as officials responded to the alert as a high-risk traffic stop, NBC first reported. “Parents, do you know where your teens are?” The San Mateo Police Department said on social media. “Waymo does!” Fast Company reached out to Waymo but did not immediately receive comment. Waymo spotted the suspicious behavior via the interior cameras and microphones aboard all Waymo robotaxis, which serve as “the ‘eyes and ears’ of our Waymo Driver,” the company’s support page says. The technology is used to make sure that vehicles remain clean, help passengers find lost items, and provide assistance during emergency situations. Waymo spotted behavior that appeared to be the teenagers shooting a weapon from the car windows, though police later determined it to be an Orbeez gun. “While there was some ingenuity to this scheme, toy guns, water guns, and BB guns all pose real dangers, especially to the untrained eye,” the San Mateo Police Department posted on Facebook. “The simple handling of them can cause fear in passersby or to those who don’t get a good look.” Additionally, law enforcement noted that even soft objects like Orbeez water beads can cause damage when projected at a high speed. Many praised the company for taking quick action, but others on social media think the reaction might not have warranted such a response. “Incredibly dangerous over-aggressive enforcement. It’s dumb kids drinking,” one person commented. And some called out the dangers of an autonomous vehicle taking the decision to lock passengers in the car without their consent. “Kids should have called the cops on Waymo for kidnapping and trafficking,” another user added on X . But for others, the bizarre event is just the beginning as robotaxi services continue to make their way into daily life. A user posted on X : “The wild part is that this probably won’t be the last strange scenario companies have to solve as self-driving vehicles become more common.”
- Everyone's Freaking Out About Microsoft's Leaked AI Windows. They Shouldn't Be
Everyone's Freaking Out About Microsoft's Leaked AI Windows. They Shouldn't Be PCMag
Score: 61🌐 MovesJul 8, 2026https://www.pcmag.com/news/everyones-freaking-out-about-microsofts-leaked-ai-windows-they-shouldnt - AI’s productivity gains are years away, but failing to deliver could make debt levels even worse
AI’s productivity gains are years away, but failing to deliver could make debt levels even worse Fortune
Score: 60🌐 MovesJul 8, 2026https://fortune.com/2026/07/08/ais-productivity-gains-jim-reid-deutsche-bank-debt/ - Reform apologises for posting fake AI image of village clean-up
Reform apologises for posting fake AI image of village clean-up The Telegraph
Score: 60🌐 MovesJul 8, 2026https://www.telegraph.co.uk/politics/2026/07/08/reform-apologises-for-fake-ai-image-sarre-clean-up-kent/ - Anthropic’s Political Risks Are Real, but OpenAI’s Loom Even Larger
The U.S. government’s targeting of Anthropic might not last—especially if political winds shift.
- How We Cut Inference Spend by Right-Sizing Our Models
Most enterprise work needs the right intelligence for each job — not the most intelligence for all of them.
Score: 60🌐 MovesJul 8, 2026https://www.salesforce.com/news/stories/cutting-inference-spend-by-right-sizing-models/ - Amazon is reportedly working on a more powerful agentic Alexa assistant
Amazon is working on a project codenamed 'Moonraker' that would enable the AI assistant to handle multi-step tasks more effectively.
Score: 60🌐 MovesJul 8, 2026https://www.engadget.com/2210614/amazon-is-reportedly-working-on-a-more-powerful-agentic-alexa-assistant/ - This Former DeepMind Exec Thinks the AI Arms Race Could End in Disaster
Verity Harding tells WIRED that the US government’s nationalistic attitude toward AI is evidence that a worst-case scenario is taking shape.
Score: 60🌐 MovesJul 8, 2026https://www.wired.com/story/verity-harding-ai-arms-race-dangers-anthology/ - From AI pilots to autonomous lending: Inside L&T Finance’s enterprise AI journey
In an exclusive interaction with Express Computer, Dr. Debarag Banerjee, Chief AI & Data Officer, L&T Finance discusses how L&T Finance is building an AI-native lending organisation, the principles behind its flagship AI platforms, and why the future of financial services will increasingly be driven by autonomous, yet responsibly governed, AI systems. The post From AI pilots to autonomous lending: Inside L&T Finance’s enterprise AI journey appeared first on Express Computer .
- Switzerland against global over-regulation of AI, says minister
Switzerland is in favour of international governance of artificial intelligence (AI), but of a “minimal” kind, the country’s communications minister has said. Albert Rösti also said he wants to see “concrete results” at a global AI summit planned for Geneva next June. “It is up to the UN to establish a minimum framework for the governance” of AI, Rösti told journalists in Geneva following a Global Dialogue on AI Governance, held on Monday and Tuesday. Above all, he believes the sector will need to introduce a certification scheme for companies based on best practice. There should be “not too much” regulation, Rösti added. In his view, the June 2027 summit must showcase Switzerland as a “hub of innovation that can provide solutions for everyone”. Switzerland will be the fifth country to host the summit, following the inaugural summit in the UK, then South Korea, France and India. As a neutral intermediary, Switzerland “can offer sound advice” whilst remaining aware of its smaller ...