AI News Archive: May 27, 2026 — Part 7
Sourced from 500+ daily AI sources, scored by relevance.
- How is A.I. affecting the job search process? | On The Coast | On Demand | CBC Listen
How is A.I. affecting the job search process? | On The Coast | On Demand | CBC Listen CBC
Score: 40🌐 MovesMay 27, 2026https://www.cbc.ca/listen/live-radio/1-46-on-the-coast/clip/16216963-how-a.i.-affecting-job-search-process - Agora, the Denodo Cloud Service, now available on the Microsoft Marketplace to Power Agentic AI Use Cases
Agora, the Denodo Cloud Service, now available on the Microsoft Marketplace to Power Agentic AI Use Cases Toronto Star
- Sonny reacts to Atlas learning football
Sonny reacts to Atlas learning football
Score: 40🌐 MovesMay 27, 2026https://www.hyundaimotorgroup.com/en/tv/sonny-reacts-to-atlas-learning-football - When GPU utilization lies: The FinOps blind spot in secure AI training
Enterprise cloud teams are trained to act on utilization data. If a virtual machine is idle, resize it. If storage is overallocated, reclaim it. If a GPU appears underused, move the job to a smaller instance. That logic is central to modern FinOps. It helps organizations reduce waste, improve forecasting and keep cloud spending under control. But secure AI training introduces a different problem: sometimes the utilization signal is technically true and operationally misleading. A GPU can look underused even when the workload is not over-provisioned. In privacy-preserving machine learning, low accelerator utilization may indicate a memory-bound bottleneck, not excess capacity. If a cloud optimization process treats that signal as ordinary waste, the recommended fix can make the job slower and more expensive. For CIOs, this is not just a GPU tuning issue. It is a cloud governance issue. As I have noted previously, IT leaders must look beyond the cloud bill to understand the hidden operational costs of AI governance The utilization number does not explain the bottleneck Traditional cloud right-sizing depends on a simple assumption: low utilization usually means unused capacity. That assumption works for many enterprise workloads. It can work for web services, batch jobs, databases and standard compute jobs. But secure AI training can break that assumption because the workload shape changes. In my IEEE systems research on privacy and robustness in machine learning , I profiled what happens when trust controls are added to model training. The important lesson for CIOs was not only that secure training costs more, but it was that secure training can change what infrastructure metrics mean. On a controlled NVIDIA V100 GPU setup, privacy-preserving training increased cost by 3.55x on a vision workload and 2.96x on a tabular workload. Robustness training increased cost by 4.07x on the vision workload. Those cost multipliers matter. But for FinOps teams, the deeper finding is this: The workload became less aligned with the hardware signals that cloud teams often use for rightsizing. Why privacy-preserving training can look inefficient Modern AI accelerators are very good at large, dense mathematical operations. Standard model training often keeps these accelerator units busy because the work can be organized into large blocks of computation. Differential privacy training often requires per-example gradient computation and clipping. Instead of pushing most of the work through large, efficient operations, the system performs more fine-grained steps across individual training examples. That changes the performance profile. In my study, this pattern created memory-bound behavior and reduced effective use of specialized GPU compute units such as Tensor Cores. To a dashboard, that can look like underutilization. To a systems engineer, it means something more specific: the job is not waiting because the GPU is too large. It is waiting because the workload is constrained by memory movement and per-example operations, simply those are not the same problem. The FinOps risk: Right answer, wrong context Automated cloud recommenders are useful because they identify resources that appear oversized or idle. The problem is not that these tools exist. The problem is applying a generic right-sizing rule to a specialized AI workload. A standard recommendation workflow might ask, “Is the accelerator busy?: For secure AI training, CIOs need the team to ask, “Why is the accelerator not busy?” If the answer is idle capacity, downsizing may save money. If the answer is memory-bound privacy computation, downsizing may increase total cost. A smaller instance may have a lower hourly price, but cloud bills are not based only on hourly price. They are based on hourly price multiplied by runtime. If the smaller instance extends the training job enough, the total bill can rise. That is the FinOps blind spot: a recommendation can look correct on a utilization dashboard but fail when measured against the full training job. Secure AI needs a different exception policy Enterprise IT already treats some workloads differently. Regulated databases, security-sensitive systems and latency-critical applications often have special infrastructure policies. Secure AI training needs similar exception handling; a model training job that uses differential privacy or adversarial training should not be evaluated the same way as an idle development server. These workloads can produce unusual utilization patterns because the algorithm itself changes the way hardware is used. 1. Tag secure-AI training jobs FinOps teams need to know when a training job uses privacy-preserving or robustness-oriented methods. A simple workload tag can prevent the job from being evaluated as ordinary compute. The tag should tell cloud teams: Low utilization may be caused by the algorithm, not by waste. This gives FinOps, MLOps and infrastructure teams a shared signal before any right-sizing decision is made. 2. Treat rightsizing as a review trigger, not an automatic action For secure AI jobs, an automated recommendation should start an investigation. It should not automatically become a change request. Before moving the workload to a smaller instance, the team should answer four questions: Is the workload compute-bound or memory-bound? Is the bottleneck caused by data loading, memory bandwidth or per-example privacy operations? Would the smaller instance reduce total job cost, or only reduce hourly rate? Has the team measured runtime impact before approving the change? This shifts FinOps from simple utilization management to workload-aware cost governance. 3. Bring MLOps into FinOps decisions FinOps teams understand pricing, commitment plans, chargeback and utilization. But secure AI workloads require another layer of interpretation. Someone must understand what the training algorithm is doing. DP-SGD and PGD do not merely consume more GPU time. They change the computation pattern. That means utilization percentage alone is not enough to make an infrastructure decision. CIOs should connect FinOps, MLOps, AI governance and infrastructure engineering before applying cost recommendations to secure AI training workloads. 4. Measure total job economics, not only instance utilization The cheapest instance is not always the lowest-cost option. For secure AI training, CIOs should require teams to compare: Hourly cost Total runtime Energy use Job completion time Model utility impact Infrastructure bottleneck profile To truly optimize these economics, teams must look beyond the hardware and apply model-level deep cuts to slash AI training costs . Ultimately, a GPU that looks underused may still be the better economic choice if it completes the workload faster and avoids a longer memory-bound run. Failing to account for the model utility impact during these infrastructure changes can easily lead organizations into the AI accuracy trap , where cost savings inadvertently ruin real-world performance. The CIO takeaway The next phase of enterprise AI will require more than model accuracy and fast experimentation. Organizations will need AI systems that are private, robust, governable and economically sustainable. In ordinary cloud operations, low utilization often means waste. In secure AI training, low utilization may mean the workload has exposed a hardware-software mismatch. The rule for CIOs is simple: Do not right-size secure AI training jobs until you understand why the accelerator is underused. In trustworthy AI, utilization is not always truth. This article is published as part of the Foundry Expert Contributor Network. Want to join?
Score: 40🌐 MovesMay 27, 2026https://www.cio.com/article/4176754/when-gpu-utilization-lies-the-finops-blind-spot-in-secure-ai-training.html - Why single-player AI is holding back the agentic enterprise
Why context and infrastructure are crucial to the agentic enterprise
Score: 40🌐 MovesMay 27, 2026https://www.techradar.com/pro/why-single-player-ai-is-holding-back-the-agentic-enterprise - Build vs. Buy AI Agents in Enterprise Applications
Build vs. Buy AI Agents in Enterprise Applications Gartner
- How to Build Reliable AI Agents Out of Unreliable AI Parts?
How to Build Reliable AI Agents Out of Unreliable AI Parts? Gartner
- Warp’s big bet on building open source with GPT-5.5
Warp uses GPT-5.5 and OpenAI models to coordinate coding agents across local, cloud, and open-source development workflows.
- Can AI really be conscious? Researchers call for more rigorous scientific standards
As artificial intelligence systems become increasingly sophisticated, questions once confined to philosophy are rapidly entering mainstream scientific and public debate: Can AI possess consciousness? Could animals, organoids, or even fetuses have subjective experiences?
Score: 40🌐 MovesMay 27, 2026https://techxplore.com/news/2026-05-ai-conscious-rigorous-scientific-standards.html - Daily Update: Viaim Raises RMB 100M; Manulife Issues S$500M Notes; Global Mofy AI Raises $8M; Ming Shing Acquires PMA Tech for $110M
Daily Update: Viaim Raises RMB 100M; Manulife Issues S$500M Notes; Global Mofy AI Raises $8M; Ming Shing Acquires PMA Tech for $110M apac.entrepreneur.com
- Physics-aware AI generates more realistic sounds by estimating mass and velocity from video
When people watch a scene in the film "Jurassic Park" where a giant dinosaur walks toward them, they naturally imagine a heavy, rumbling sound, as if the ground were shaking. This is because humans predict sound by considering not only the shape of an object, but also physical properties such as its size, weight, and speed of movement. However, existing video-to-audio generation AI mainly generates sound based on the category of objects or scene information in the video, and has not sufficiently reflected physical properties that vary depending on weight or speed.
Score: 39🌐 MovesMay 27, 2026https://techxplore.com/news/2026-05-physics-aware-ai-generates-realistic.html - Why your company's stored data probably isn't ready for AI
Most AI initiatives fail not because the models are wrong but because the data beneath them was never prepared for the job
- IAB Tech Lab tackles the growing AI bot problem
New guidance from IAB Tech Lab aims to help publishers and content owners decide how to manage AI crawlers, bots, and other non-human traffic. The post IAB Tech Lab tackles the growing AI bot problem appeared first on MarTech .
- ETRI develops digital twin-based software for wearable robot evaluation
ETRI develops digital twin-based software for wearable robot evaluation EurekAlert!
- MEMO: A Modular Framework for Training a Dedicated Memory Model on New Knowledge Without Modifying LLM Parameters
MEMO: A Modular Framework for Training a Dedicated Memory Model on New Knowledge Without Modifying LLM Parameters MarkTechPost
- HCLTech Launches Autonomous Finance Platform Built on Google Cloud’s Gemini Enterprise
Discover HCLTech's autonomous finance platform built on Google Cloud's Gemini Enterprise AI. Learn how it automates workflows, reduces manual intervention, and
- Hurricane air-sea drag saturation and sea-state dependence revealed by surface drones
Science Advances, Volume 12, Issue 22, May 2026.
- Capgemini sets new targets building on AI consulting, but shares fall
Capgemini sets new targets building on AI consulting, but shares fall Reuters
Score: 39🌐 MovesMay 27, 2026https://www.reuters.com/business/capgemini-sets-mid-term-targets-it-bets-ai-consulting-growth-2026-05-27/ - Are the chemicals around you safe? Researchers are using AI to find out
People are exposed to thousands of chemicals every day—through the products they use, the food they eat and the environments they live in—but only a fraction of those chemicals have been fully tested for safety.
- Xage extends zero trust to autonomous AI agents across cloud, SaaS and edge
Zero-trust cybersecurity company Xage Security Inc. today unveiled new capabilities in its platform designed to give enterprises deterministic visibility into autonomous artificial intelligence agents and block them from taking unauthorized actions when compromised. The release introduces two components, Xage Agent Sentry and Xage Resource Gateway, that the company says together can wrap an AI agent […] The post Xage extends zero trust to autonomous AI agents across cloud, SaaS and edge appeared first on SiliconANGLE .
Score: 39🌐 MovesMay 27, 2026https://siliconangle.com/2026/05/27/xage-extends-zero-trust-autonomous-ai-agents-across-cloud-saas-edge/ - Preparing for the Autonomous Era in ITOps
Preparing for the Autonomous Era in ITOps IT Pro
Score: 39🌐 MovesMay 27, 2026https://www.itpro.com/technology/preparing-for-the-autonomous-era-in-itops - PlainID Expands Support for AWS AgentCore and Microsoft Foundry Across Its Agentic Security and Authorization Platform
PlainID Expands Support for AWS AgentCore and Microsoft Foundry Across Its Agentic Security and Authorization Platform azcentral.com and The Arizona Republic
- TIME Brings Together Influential Leaders for First-Ever TIME100 AI Leadership Forum
TIME Brings Together Influential Leaders for First-Ever TIME100 AI Leadership Forum Time Magazine
- Inspired by armadillos, this soft robotic shell flips from flexible to fortress in an instant
Researchers have drawn inspiration from armadillos to create a protective structure that responds to external threats by curling into a protective ball to protect electronic devices or other payloads. The structure is designed to automatically respond when it detects strain and can be tuned to respond to anything from a delicate touch to a significant impact.
Score: 38🌐 MovesMay 27, 2026https://techxplore.com/news/2026-05-armadillos-soft-robotic-shell-flips.html - e-con Systems to Showcase AI-Powered Smart Cameras and Urban Mobility Vision Solutions at ITS America Conference & Expo 2026
e-con Systems to Showcase AI-Powered Smart Cameras and Urban Mobility Vision Solutions at ITS America Conference & Expo 2026 Toronto Star
- From insights to execution: How advanced AI agents drive improved enterprise decision-making
From insights to execution: How advanced AI agents drive improved enterprise decision-making Techcircle
- Doppel launches agentic email security to disrupt phishing campaigns at the source
Social engineering defense startup Doppel Inc. today launched Doppel Email Security, an agentic artificial intelligence layer that traces phishing messages back to attacker infrastructure and orchestrates takedowns of the broader campaign rather than quarantining individual emails. The product joins Doppel’s existing Digital Risk Protection and Human Risk Management offerings on a unified platform aimed at […] The post Doppel launches agentic email security to disrupt phishing campaigns at the source appeared first on SiliconANGLE .
Score: 38🌐 MovesMay 27, 2026https://siliconangle.com/2026/05/27/doppel-launches-agentic-email-security-disrupt-phishing-campaigns-source/ - NVIDIA Dynamo Snapshot: Fast Startup for Inference Workloads on Kubernetes
The cold-start problem In production inference deployments, demand fluctuates over time, requiring inference replicas to scale elastically. However,...
Score: 38🌐 MovesMay 27, 2026https://developer.nvidia.com/blog/nvidia-dynamo-snapshot-fast-startup-for-inference-workloads-on-kubernetes/ - Infortrend to Showcase Next-Generation AI Infrastructure from Edge to Cloud at COMPUTEX 2026
Infortrend to Showcase Next-Generation AI Infrastructure from Edge to Cloud at COMPUTEX 2026 The Straits Times
- Apacer Showcases Edge AI Storage Power at COMPUTEX 2026
Apacer Showcases Edge AI Storage Power at COMPUTEX 2026 The Straits Times
- Sungsimdang's famous 'soboro' bread may soon be baked by AI-powered machines, not human hands
Fried "soboro" bread, a type of sweet bun [JOONGANG ILBO] Sungsimdang's famed fried soboro bread, a type of sweet bun with a crunchy exterior and a soft, buttery filling, may soon be baked by AI-powered machines instead of human hands. Sungsimdang, a popular bakery synonymous with Daejeon, plans to automate the entire production process of its fried soboro bread — from dough preparation and baking to packaging — using AI and robotics. Related Article Pope Leo XIV gives blessings to Daejeon bakery Sungsimdang for 70th anniversary No soboro for you: Iconic Daejeon bakery Sungsimdang to close Nov. 3 for employee sports festival Daejeon bakery Sungsimdang beats major franchises with record sales in 2024 Founded in 1956 as a steamed bun shop near Daejeon Station, Sungsimdang has grown into a symbol of the city. It is famous for its fried soboro bread and “siru” cakes, which are layered with generous amounts of fruit, including strawberries, figs and mangoes, depending on the season. The bakery expects the system to boost productivity by roughly 20 percent, easing the workload of employees who have long carried out repetitive tasks in high-temperature environments. Industry Minister Kim Jung-kwan visited a Sungsimdang branch at a Lotte Department Store in the city on Wednesday to inspect a pilot program of an AI-powered smart factory system. He also discussed plans to accelerate the expansion of AI manufacturing technologies. An automated AI-powered process for baking and preparing fried "soboro" bread, a type of sweet bun [MINISTRY OF TRADE, INDUSTRY AND RESOURCES] The government, which initially focused its AI transformation initiative on manufacturing sectors such as semiconductors, automobiles and shipbuilding, has broadened the push into everyday industries, including food production and services, since last year. A total of 102 AI-powered factories have been established, with another 100 slated to roll out this year. “We confirmed that the AI used to detect semiconductor defects and the AI used to identify flaws in the soboro bread are technologically very similar,” the minister said. “We will expand the initiative beyond manufacturing and across the economy.” This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom. BY JEONG JAE-HONG [lee.soojung1@joongang.co.kr]
- The Best AI Observability Tools for Agentic Systems in 2026
AI applications used to rely on a handful of straightforward LLM calls. Now agents make hundreds of decisions in response to a single user input, calling tools, retrieving context, and compounding outputs. When something goes wrong, the failure can be six steps deep and invisible from the outside. Most AI observability tools were designed to […] The post The Best AI Observability Tools for Agentic Systems in 2026 appeared first on Comet .
Score: 38🌐 MovesMay 27, 2026https://live-comet-marketing-site.pantheonsite.io/blog/ai-observability-tools/ - Industry-standard LLM benchmarks in DataRobot
Every LLM deployment has a ceiling, a latency curve, and a unit cost. Most teams operate blindly, discovering their deployment limits only when over-provisioning exhausts their GPU budget or peak traffic causes a catastrophic failure. Three numbers matter: maximum sustained concurrency before GPU saturation, end-to-end latency at that concurrency, and cost per million tokens at... The post Industry-standard LLM benchmarks in DataRobot appeared first on DataRobot .
- 10 Problems AI Can Help Fashion Solve
Here's how fashion brands are deploying artificial intelligence to tackle real business problems and work faster, smarter and more efficiently.
Score: 38🌐 MovesMay 27, 2026https://www.businessoffashion.com/articles/technology/10-problems-ai-can-help-fashion-solve/ - AI Proves Language Evolves for Learnability
A new study using AI proves that language actively evolves over generations to become structured because it makes learning easier.
Score: 38🌐 MovesMay 27, 2026https://neurosciencenews.com/ai-iterated-learning-deep-linear-networks-30770/ - Why Legacy tools around Jira quietly block enterprise AI adoption (2026)
Why Legacy tools around Jira quietly block enterprise AI adoption (2026) Atlassian Community
- Anvilogic AI SOC Platform vs LogRhythm SIEM 2026
Anvilogic AI SOC Platform vs LogRhythm SIEM 2026 Gartner
- Expecting the Unexpected: Monitoring for Drift in ML Systems
Expecting the Unexpected: Monitoring for Drift in ML Systems CMU Software Engineering Institute
Score: 38🌐 MovesMay 27, 2026https://www.sei.cmu.edu/blog/expecting-the-unexpected-monitoring-for-drift-in-ml-systems/ - Introducing the perception agent harness with annotation and verification
Today we're announcing the open-source release of the first two primitives for our perception agent harness: annotation and verification.
- Meet EAGLE 3.1: The Speculative Decoding Algorithm That Fixes Attention Drift in LLM Inference
Meet EAGLE 3.1: The Speculative Decoding Algorithm That Fixes Attention Drift in LLM Inference MarkTechPost
- AI Labs: Elon Musk wants AI in space
Will SpaceX’s giant IPO turbocharge Elon Musk’s AI ambitions?
- The case for AI-powered work management in a growing business
The use of AI in work management is on the rise, with more than a third of workers now using AI tools daily, according to recent data. In fact, the introduction of generative tools like ChatGPT and other major AI-powered work management platforms has enabled 53 per cent of users to work more efficiently, 48 […] The post The case for AI-powered work management in a growing business appeared first on e27 .
Score: 38🌐 MovesMay 27, 2026https://e27.co/the-case-for-ai-powered-work-management-in-a-growing-business-20260524/ - Google's CEO Had Notes When Shown 'Opinionated' AI Search Results
Google's CEO Had Notes When Shown 'Opinionated' AI Search Results Business Insider
Score: 38🌐 MovesMay 27, 2026https://www.businessinsider.com/google-ceo-opinionated-search-results-ai-overview-sundar-pichai-2026-5 - Why Memory Chips Are Dominating the A.I. Rally
Three major producers — Micron, Samsung and SK Hynix — are now trillion-dollar companies. Politicians and Wall Street have taken notice.
Score: 38🌐 MovesMay 27, 2026https://www.nytimes.com/2026/05/27/business/dealbook/ai-chips-war-samsung-micron.html - Micron Stock: The ‘Insatiable’ Logic Behind the Memory Maker’s ‘Extreme’ Gains.
Micron Stock: The ‘Insatiable’ Logic Behind the Memory Maker’s ‘Extreme’ Gains. Barron's
Score: 38🌐 MovesMay 27, 2026https://www.barrons.com/articles/micron-stock-price-memory-chip-sk-hynix-049d3a30 - UC Berkeley bans AI use for law students
The school prohibits students from using AI to complete assignments, brainstorm ideas, outline papers, and even correct grammar.
Score: 37🌐 MovesMay 27, 2026https://www.semafor.com/article/05/27/2026/uc-berkley-bans-ai-use-for-law-students - Mphasis launches autonomous enterprise agency platform, returns expected by FY28
The company invested 1.5 per cent of its FY26 revenue ($27 million), betting on FY27 as the foundational layer and expecting returns FY28 onwards.
- Anthropic strengthens India team, appoints Sangeeta Bavi to lead startups
Announcing the move in a post on LinkedIn, Bavi pointed to her long ties with the founder community. "Founders have always held a special place in (my heart)," she wrote. "Their ambition, resilience and belief that they can change the world have taught me more than any other form of learning. That’s why joining Anthropic as Head of Digital Natives, Startups & Growth, India feels deeply meaningful."
- LightTable Raises $22 Million Series A to Accelerate AI-Native Workflows for Design and Construction Teams
LightTable Raises $22 Million Series A to Accelerate AI-Native Workflows for Design and Construction Teams markets.businessinsider.com
- From Search Engines to Autonomous Agents: AI industry enters its next phase
The AI industry is rapidly shifting from conversational chatbots to autonomous “Agentic AI” systems capable of independently executing complex tasks across workflows and applications. Companies including OpenAI, Microsoft, Google, and Salesforce are leading investments in AI agents designed to automate research, customer support, scheduling, and enterprise operations. The shift is expected to redefine productivity, workplace automation, and how businesses interact with software globally.