AI News Archive: May 26, 2026 — Part 4
Sourced from 500+ daily AI sources, scored by relevance.
- Young people chasing jobs in tech and professional services hit hardest by AI – conference hears
Department of Finance economists tells Neri conference that impact of AI on labour markets remains uncertain
- Anthropics Claude Mythos, or a model like it, to get public release
Anthropic is developing guardrails to prepare its Mythos AI model for public release.
- Hyundai taps former Nvidia engineer to boost self-driving, AI
Hyundai Motor Group has hired former Nvidia autonomous driving expert Lee Hee-seok to its software research unit 42dot, amid the South Korean automaker’s push to strengthen its software-defined vehicles and self-driving technology capabilities. According to industry sources Tuesday, 42dot appointed Lee, an expert in computer vision technology, as vice president to lead research on vision-language action models. This marks the first time an executive has been brought in from outside the company s
- This Orbital Data Center Startup Is Integrating Starlink Lasers
This Orbital Data Center Startup Is Integrating Starlink Lasers PCMag
Score: 52🌐 MovesMay 26, 2026https://www.pcmag.com/news/this-orbital-data-center-startup-is-integrating-starlink-lasers - Gartner Says Applying Uniform Governance Across AI Agents Will Lead to Enterprise AI Agent Failure
Gartner Says Applying Uniform Governance Across AI Agents Will Lead to Enterprise AI Agent Failure Gartner
- Exclusive research: Banks up AI investment to cut costs
Banks are prioritizing AI investments for several reasons, including streamlining operations, reducing costs and improving employee workflows. By comparison, enhancing risk assessment and strengthening cybersecurity ranked lower among the reasons banks are investing in AI.
Score: 52🌐 MovesMay 26, 2026https://www.americanbanker.com/creditunions/news/exclusive-research-banks-up-ai-investment-to-cut-costs - Mistral AI Takes Aim at Legal Sector Through Expanded Harvey AI Partnership
The French AI company is bringing its models to the legal sector under a partnership with Harvey AI, taking aim at a lucrative industry where rivals like Anthropic are expanding aggressively.
- Exclusive: Ex-Palantir AI execs raise $12 million seed round for Perceptic, a startup automating drug discovery
Exclusive: Ex-Palantir AI execs raise $12 million seed round for Perceptic, a startup automating drug discovery Fortune
- Inside Human Archive: The Y Combinator Startup Recording Indian Workers To Train The World’s Robots
Human Archive is making waves in India’s startup ecosystem this week amid reports of its work with home services startups…
Score: 52🌐 MovesMay 26, 2026https://inc42.com/startups/inside-human-archive-the-startup-recording-indian-workers-to-train-the-worlds-robots/ - WhatsApp is bringing document sharing support to Meta AI: What it means
WhatsApp is reportedly testing document-sharing support for Meta AI on Android, allowing beta users to upload files directly for analysis, summaries, and assistance within chats
- As agentic AI surges, CPUs and air-cooled infrastructure move to the fore
Agents are turning up the heat for enterprises, turning air-cooled AI infrastructure into a boardroom priority. While GPUs have largely dominated the AI conversation, CPUs are increasingly coming into focus. Agentic AI relies on constant orchestration, data processing, and decision-making between tasks — work that plays to the CPU’s strengths, not the GPU’s. The math alone […] The post As agentic AI surges, CPUs and air-cooled infrastructure move to the fore appeared first on SiliconANGLE .
Score: 51🌐 MovesMay 26, 2026https://siliconangle.com/2026/05/26/air-cooled-ai-infrastructure-demand-agentic-era-delltechworld/ - AI is rewiring digital media’s power structure
AI is rewiring digital media’s power structure verdict.co.uk
Score: 51🌐 MovesMay 26, 2026https://www.verdict.co.uk/analyst-comment/ai-rewiring-digital-media-power-structure/ - AI is ‘going to break down millions of careers,’ Gartner analyst says
Organizations need to rethink how employees gain expertise, or they will find themselves without talent for the jobs artificial intelligence helps create, the analyst said.
Score: 51🌐 MovesMay 26, 2026https://www.hrdive.com/news/ai-going-to-break-down-millions-of-careers-gartner/821065/ - Every organization is pouring money into AI right now, and almost none of them know what their people are actually doing with it': Study reveals employees are using their personal AI accounts at work, raising a whole host of issues
Shadow AI exposes company data and risks IP loss – but personal tools offer easier access than clunky enterprise accounts.
- LimX Dynamics unveils Luna humanoid robot with AI dance learning
LimX Dynamics on Monday unveiled the LimX Luna humanoid robot, priced at RMB 298,000 ($41,000). Standing 160cm tall, the LimX Luna features 27 degrees of freedom across its body and is powered by the company’s second-generation SYS 0 motion control engine. The robot also comes with upgraded cooling and battery life, while supporting multimodal interaction […]
Score: 51🌐 MovesMay 26, 2026https://technode.com/2026/05/26/limx-dynamics-unveils-luna-humanoid-robot-with-ai-dance-learning/ - GovernmentAI Is Moving at the Speed of Innovation. Our Ability to Evaluate It Must Too.
A framework for evaluating AI in national security
- This startup is betting India’s gig economy can train the world’s robots
Human Archive, a startup founded by UC Berkeley and Stanford researchers, is paying gig workers in India to wear camera-equipped caps and sensor devices to collect the real-world physical training data that AI and robotics labs are racing to acquire.
- How this wearable AI technology is helping NBA, NHL and athletes everywhere prevent injuries
Gabriel Landeskog wears the small sensors in the insoles of his skates for practices and games. He wears them in his sneakers when he’s training and, maybe most handy of all, while taking his dog for a walk. Those spins around the block and ice record all of his biomechanical measurements . The numbers provided a blueprint in helping the Colorado Avalanche captain resume his career after a three-year gap caused by a complicated knee injury. Now, they keep him at his gritty, goal-scoring best. The collected data ranges from movement patterns to his asymmetry and whether he’s favoring his surgically repaired right knee. It calculates in-game/in-practice workloads, stride characteristics, and the mechanics of how his feet interact with various surfaces—ground or ice. The details paint a picture to inform Landeskog when he’s reaching maximum capacity and needs a break. That way, the technology prevents him from reaching overexertion levels in training that might set him back for days, possibly weeks. “This detects any red flags before I even feel them,” said Landeskog, whose team trails Vegas 3-0 in a Western Conference Final in which he has two of the Avalanche’s six goals. “It’s been super important for me, and a huge help.” The assist goes to Plantiga, an AI-driven movement platform that helps athletes stay on top of their game and prevent injuries. The company’s cutting-edge technology is being utilized by players and teams in the NBA, NFL, WNBA and MLB, along with colleges, elite sprinters, weekend warriors and, of course, NHL players such as Landeskog. “What we’re trying to detect is the smoke before the fire,” explained Matthew Jordan, the vice president of performance science at Plantiga as well as an associate professor, faculty of kinesiology/sport medicine center, at the University of Calgary. “Imagine you’re at the point where your knee is just at the cusp of the next day it’s going to be like, ‘My knee’s killing me. I can hardly walk.’ We can see in the data before you reach that tipping point.” Landeskog’s return from knee injury Landeskog’s knee issues began after a skate blade cut his right knee during the 2020 playoffs in the Edmonton pandemic bubble. He worked his way through it and helped the Avalanche to a Stanley Cup title in 2022 by beating Tampa Bay. That Cup clincher, though, was his last game for a while. After missing a full season, Landeskog underwent cartilage replacement surgery on May 10, 2023. Introduction to Plantiga In the spring of 2024, Landeskog was introduced to Plantiga, the Vancouver-based human analytics company founded by Quin Sandler and his late father, Norman McKay. They wanted to create a way to monitor the movement of athletes with wearable in-shoe technology. Landeskog reached out to the Plantiga team through the strength and conditioning coach Marcin Goszczynski. The 33-year-old Landeskog met with Jordan at a game when the Avalanche were playing in Calgary. “We discussed his injury and his frustration with the process,” Jordan recounted. “You have to remember at this point the tunnel was dark and long—there was no light . . . we were miles from the end of the tunnel.” Jordan connected Landeskog with a Canadian ski racer who went through a similar injury. “It was a relief for Gabe to know that another athlete out there had been able to conquer this injury,” Jordan said. “He has among the best mindsets, and he is 100% resilient and gritty to the core.” By utilizing Norman, a movement intelligence layer named after Sandler’s father, potential changes in Landeskog’s biomechanics were flagged before they could escalate. “We’re trying to put really good data [together] that him and his trainer will use,” Sandler said. “There’s this fine Goldilocks zone that we help him stay in, and honestly he’s been killing it.” Staying on top of the data Landeskog returned last season for Game 3 of the playoffs against Dallas, which was his first NHL contest in 1,032 days. His comeback continued this season, when he had 14 goals and 21 assists over 60 regular-season games. Throughout the season, Jordan tunes in to watch Landeskog’s strides on the ice. He sometimes notes instances he wants to examine further simply because the numbers may be outside the Swedish forward’s normal range. “Essentially, put out the ‘smoke’ before it turns into a ‘fire,'” Jordan explained. “In an athlete’s world, a fire can mean a new injury, a reinjury to the tissue, a loss of performance, or a setback in rehab.” This application is similar to the Oura Ring, which constantly collects health and wellness metrics. Plantiga, though, tracks human movement through a laboratory-grade inertial measurement unit sensor that captures 400 data points per second. Translation: An athlete’s movement can be captured with 20 to 30 times more granularity than with a smart phone or watch. “A supercharged human movement measuring device,” Jordan said. One way to get a baseline for Landeskog’s gait and biomechanics was through walking. For that, an assist goes to his dogs, the late Zoey and now Mila, who were eager participants on those data-collecting excursions. “We can see subtle things in your walk patterns well before it manifests as something very clinical or significant,” Jordan said. Taking out the guesswork What the data did for Landeskog was take the guesswork out of his training program. “He’d get on the ice and be like, ‘Oh, I feel good today.’ Jordan said. “It’s like, ‘I think I’m just going to go hard. I feel like my knee feels really good. Oh [no], I went too far. My knee’s flared up. I’ve got to take a week off.’ With all these setbacks he couldn’t catch any progression.” Now, when the numbers indicate Landeskog should rest, he pays attention. He’s a finalist for the Bill Masterton Memorial Trophy, which is awarded to the NHL player who exemplifies perseverance, sportsmanship, and dedication. “I’m humbled and honored by it, but I think for me, the ultimate prize I’ve already won,” Landeskog said. “That’s to continue working and getting to play hockey.” AP NHL: https://apnews.com/hub/stanley-cup and https://apnews.com/hub/nhl —Pat Graham and Stephen Whyno, AP Sports Writers
- Jensen Huang says parents shouldn’t worry about what their kids study as China cuts non-AI degrees
Jensen Huang says parents shouldn’t worry about what their kids study as China cuts non-AI degrees Fortune
Score: 51🌐 MovesMay 26, 2026https://fortune.com/2026/05/26/jensen-huang-nvidia-ai-colleges-and-universities-china-cutting-arts-degrees/ - Didit raises $6M funding to build AI-native identity infrastructure
Artificial intelligence-native infrastructure provider for identity and fraud prevention Didit today announced it has raised $6 million in seed funding. New and continuing investors joining the round included Y Combinator LLC, Pioneer Fund, Orange Collective, Founders Future, Phosphor Capital, SaaSholic and Rebel Fund, alongside angel investors of Gusto Inc. Co-founder Tomer London, Fond Technologies Inc. […] The post Didit raises $6M funding to build AI-native identity infrastructure appeared first on SiliconANGLE .
Score: 50💰 MoneyMay 26, 2026https://siliconangle.com/2026/05/26/didit-raises-6m-funding-build-ai-native-identity-infrastructure/ - Could AI Run Your Airport Security Line? Inside TSA’s New Gold+ Initiative
Private partners and new technology could soon reshape airport security operations across the country.
- Government sets target for ‘revised’ AI policy
The finalised AI policy is still due by the end of the financial year, after the draft was withdrawn due to “fictitious” AI input.
Score: 50🌐 MovesMay 26, 2026https://www.itweb.co.za/article/government-sets-target-for-revised-ai-policy/lwrKxv3Y2QyMmg1o - What Upwork, DoorDash, Meta, EY, and Fundrise reveal about agents
Subscribe • Previous Issues Beyond the Demo: What Real AI Agents Actually Do at Work I am always on the lookout for new AI agents and applications that operate outside the coding world. By agent, I mean a system that can take a goal, use tools, keep context, and work through several steps rather than simply answer Continue reading "What Upwork, DoorDash, Meta, EY, and Fundrise reveal about agents" The post What Upwork, DoorDash, Meta, EY, and Fundrise reveal about agents appeared first on Gradient Flow .
Score: 50🌐 MovesMay 26, 2026https://gradientflow.com/ai-agents-outside-the-coding-world-and-what-makes-them-stick/ - From Process to Chip: How Wipro and Intel Are Delivering ROI-First AI for the Enterprise
AI has already begun transforming several Industries, business models & processes. Accordingly, Enterprises have been pivoting their strategy to ensure they are well positioned to leverage AI for Growth, higher Margins and Efficiency. However, the ROI of AI initiatives is as much a function of the execution model as it is of the Use case, since AI is also upending the technology stack that delivers the Enterprise & Process transformations. The key challenge in execution is integrating the various components in the technology stack (hardware, software, domain knowledge, Services, Security etc.) in a cost-efficient manner without compromising on scale and performance. A Partnership Built Around the Enterprise Pain Point The partnership between Wipro and Intel addresses exactly this challenge — providing enterprise clients with an integrated solution that co-innovates by combining Intel’s expertise in chip design, compute efficiency, and security with Wipro’s strengths in domain knowledge, process mapping, data management, and analytics & AI. The partnership between Wipro and Intel addresses exactly this pain point – providing Enterprise clients with an integrated solution that co-innovates based on Intel’s expertise on chip design, compute efficiency, security with Wipro’s expertise in Domain, process mapping, data management and Analytics & AI. Pushpa Ramachandran – Vice President, Head of Data & Decision Sciences at Wipro explains that the philosophy is to deliver solutions optimized for scale, cost and time by adding value at every step, from ‘Process to Chip’ and everything in between. Solutions that have been jointly built include a re-engineered Loan origination system, an AI & Gen AI led SAR (Suspicious Activity Reporting) solution, automated & tailored Marketing campaign Manager and several Computer Vision use cases. These Solutions are designed to deliver ROI based on the philosophy of having the most optimal solution for every problem. Tackling TCO, Sovereign AI, and Responsible AI Raghavendra Ural, Director of Engineering, Datacenter AI Solutions at Intel further adds that optimized solutions work on 3 key client requirements of TCO, Sovereign AI and Responsible AI. By leveraging SLMs (rather than LLMs) that suffice for most Use cases and using Intel XEON CPUs for model inferencing, TCO can be brought down by up to 2/3 rd of potential costs. Besides, the design allows for model & Solution deployment in Private clusters to maintain privacy and security, in addition to Intel’s own embedded security layers such as TDX (Trust Domain extensions) and SGX (Software guard extensions). This optimized design ensures much lower carbon footprint than other architectures that over-provision the compute (GPU) capacity required. Speed, Agility, and the VEGA Accelerator It is important to note that offering an AI first, optimized & integrated solution that delivers ROI is critical but not sufficient. Enterprise clients also demand Speed of implementation while emphasizing the need for AI to be Sustainable and Responsible. Wipro addresses the Speed and Agility requirement through VEGA, WIPRO’s accelerator that enables Enterprises to deploy these solutions based on internal data quickly and then also monitor the efficacy of the solutions over time. AI models often suffer from Model ‘drift’ and hallucinations which VEGA helps to track and rectify. Sustainability: Designing for Minimal Carbon Footprint In recent times, Enterprises have sharpened their focus on minimizing the Carbon footprint in their business. AI compute clusters can be Water and Power intensive. This is where Intel’s multi-decade experience in optimizing tech stacks comes in handy as these Solutions can be designed for minimal carbon footprint impact. Raghavendra points out that Industry benchmark metrics is seeing a shift from ‘$ per token’ to ‘Token per Watt’. Intel and Wipro believe their Solution design is best positioned to address this shift and meet the multiple needs of their Enterprise clients as Clients leverage AI for their business.
- Ignite, OST drive to solve autonomous vehicle challenges with AI
As regulatory authorities in the UK open applications for operators to run autonomous taxis, buses and private-hire cars, scalable software-defined vehicles (SDVs) venture Ignite by Forvia Hella has announced a collaboration with Samsung Electronics-owned artificial intelligence (AI) software firm Oxford Semantic Technologies (OST) to create an “explainable” AI service to help prove road safety compliance for autonomous vehicles (AVs). The partners noted that their partnership comes as leading autonomous vehicle providers face challenges as their vehicles continue to deal with complex decision-making, resulting in, in some cases, vehicles driving into flooded roads . In addition, the partners say that while driving performance is improving in AVs, manufacturers still lack robust ways to prove safety, compliance and decision logic at scale. This, they insist, has created a blocker on progressing to higher levels of autonomy, with manufacturers struggling to move from Level 2 (partial driving automation where the driver is legally responsible) to Level 3 (conditional driving automation where the manufacturer assumes liability) and Level 4 (fully autonomous driving) . The collaboration will aim to address the industry’s struggle to progress from Level 2 to Level 4 autonomy by providing ways to prove safety, compliance and decision logic at scale. It will also look to provide software engineers with a “white box” so they can understand how AI makes decisions, to improve safety features and seek approval with hard evidence. For autonomous vehicles, knowledge-based AI can be used as the vehicle’s rulebook and memory – capturing everything the car does in real time, cross-referencing it against traffic rules, and ensuring every decision it makes is logical, traceable and compliant. This provides software engineers with a white box of data they can use to better understand the AI that powers AVs, and improves safety features and performance. It also has the potential to produce evidence for regulators on how AVs make decisions in different circumstances and conditions, and follow road traffic rules. From a technological standpoint, the simulation-based software uses OST’s RDFox knowledge graph database to provide a reasoning layer to AV systems, improving decision-making in complex situations. The simulation-based software is attributed with bridging the gap between traffic laws and live autonomous decision-making. Read more about autonomous vehicles UK government accelerates autonomous vehicle development funding : Projects exploring how autonomous vehicles could benefit businesses and communities across the UK receive government backing as part of £150m CAM Pathfinder programme. Wayve gears up with end-to-end AI for autonomous vehicles : Mobile technology platform firm teams with UK self-driving company to advance production-ready end-to-end artificial intelligence for assisted and automated driving. Motive accelerates Edge AI safety for automotive operations : Commercial vehicle AI dash cam said to be able to deliver three times more AI processing power, stereo vision and hands-free two-way communication in an all-in-one device. Rubber hits the road for Qualcomm automotive : Mobile tech leader uses CES to outline advances in automotive through key collaborations with Chinese startup technology company, IT behemoth and manufacturing group to boost, ADAS, IVI and AI compute. With its AI-centric engine, RDFox OST currently collaborates with organisations across Europe, Asia and North America. From integrating data across organisations to autonomous decisions and recommendations, OST’s technology is deployed in financial services, automotive, manufacturing, healthcare, publishing and retail. The collaboration is also claimed to demonstrate how knowledge-based AI, which uses carefully curated expert knowledge and logical reasoning to solve complex problems, can be applied to the AV sector. Unlike machine learning, which finds patterns in vast datasets and draws statistical outputs, knowledge-based AI aims to improve the accuracy of results by making logical and explainable decisions based on data combined with expert knowledge. “Hella Ignite.Drive applies knowledge-based AI by translating traffic laws, originally written for human interpretation, into machine-readable rule sets,” said Felix Kortmann, chief technology officer at Ignite by Forvia Hella. “This enables manufacturers to generate deterministic evidence that demonstrates safe and compliant vehicle behaviour for European type approval. “At the same time, it reduces development lead times by minimising the need for manual, market-by-market rule coding, helping AV teams move faster toward approval-ready deployment,” he said. Oxford Semantic Technologies CEO Peter Crocker said: “Autonomous vehicles currently use AI to make a whole range of decisions on the road – but at the moment, manufacturers are struggling to show why or how these decisions are made. RDFox can help with this major barrier to progress. What knowledge-based AI allows us to do is collect and map these decisions and apply reasoning. We can see exactly why a vehicle acts in a certain way and use this data to help the vehicle make better decisions in the future.” Ian Horrocks, Oxford University professor and OST co-founder, added: “The AV case study is a great example of how knowledge-based AI can enhance data-driven systems. A key advantage of the technology is traceability, where decisions can be linked back to the rules and logic that produced them. In the automotive space, this visibility can revolutionise go-to-market strategies, improving the compliance and safety of AVs.”
Score: 50🌐 MovesMay 26, 2026https://www.computerweekly.com/news/366643508/Ignite-OST-drive-to-solve-autonomous-vehicle-challenges-with-AI - Japan cable maker rout exposes cracks in AI infrastructure rally
Japan cable maker rout exposes cracks in AI infrastructure rally The Japan Times
Score: 50🌐 MovesMay 26, 2026https://www.japantimes.co.jp/business/2026/05/26/companies/japan-ai-stocks-fragility/ - AI Agents: Cybersecurity Strategies for Enterprise Protection
AI Agents: Cybersecurity Strategies for Enterprise Protection Gartner
- Beyond the buzz: How AI and sustainability are reshaping design, manufacturing, and construction in APAC
A new report, the 2025 State of Design & Make Report – APAC, sheds light on the evolving landscape of architecture, engineering, construction and operations (AECO), design and manufacturing (D&M), and media and entertainment (M&E) industries across the Asia Pacific (APAC) region, including key Southeast Asian nations. The study, based on surveys and interviews with […] The post Beyond the buzz: How AI and sustainability are reshaping design, manufacturing, and construction in APAC appeared first on e27 .
- Has Google Already Won the A.I. Race?
Has Google Already Won the A.I. Race? Puck
- Stability AI Releases Stable Audio 3: A Family of Fast Latent Diffusion Models for Audio Generation and Editing
Stability AI Releases Stable Audio 3: A Family of Fast Latent Diffusion Models for Audio Generation and Editing MarkTechPost
- McClatchy union journalists at 4 WA newspapers strike over pay, AI use
Affected newspapers include The Tacoma News Tribune, the Tri-City Herald, The Olympian and the Bellingham Herald.
- Can you own a voice? Taylor Swift’s latest legal move raises big questions for AI and copyright
Swift appears to be the first musician to take this step.
- Lucis raises $20M Series A to expand AI-driven preventive healthcare platform
Preventive health platform Lucis has raised $20 million Series A, led by Singular, with participation from General Catalyst, Y Combinator, and angels including investors behind Runna, Céline Lazorthes...
Score: 50💰 MoneyMay 26, 2026https://tech.eu/2026/05/26/lucis-raises-20m-series-a-to-expand-ai-driven-preventive-healthcare-platform/ - FirstFT: ByteDance offers AI team special stock to combat poaching
Also in today’s newsletter: BP removes chair Albert Manifold, and Iran accuses US of ‘flagrant’ ceasefire violations
- ‘BusPatrol’ Put AI Cameras in Tens of Thousands of School Buses. Now They Want to Give Cops Access
BusPatrol plans to scan the license plates of all vehicles the buses drive past, and then let law enforcement search that data. The plan would essentially turn school buses into roaming surveillance vehicles.
- Should AIs be required to report a human user contemplating violence?
Human therapists have a legal duty to warn authorities and potential targets when patients say they plan to harm someone. The same can – in theory – be required of AI chatbots .
Score: 50🌐 MovesMay 26, 2026https://theconversation.com/should-ais-be-required-to-report-a-human-user-contemplating-violence-282561 - DeepSeek Slashes V4 Pro Prices by 75% Permanently
The company’s aggressive pricing strategy is expected to increase pressure on AI companies globally.
Score: 49🌐 MovesMay 26, 2026https://analyticsindiamag.com/ai-news/deepseek-slashes-v4-pro-prices-by-75-permanently - China's AI talent war heats up
China and its tech giants are racing to protect their top AI talent, as competition intensifies at home and with the US.
Score: 49🌐 MovesMay 26, 2026https://www.semafor.com/article/05/26/2026/chinas-ai-talent-war-heats-up - MiniCPM5-1B: The leading 1B open weights model
MiniCPM5-1B is a leading 1B open weights model released on May 26, 2026.
Score: 49🤖 ModelsMay 26, 2026https://artificialanalysis.ai/articles/minicpm5-1b-the-leading-1b-open-weights-model - Canada’s AI champion Cohere is standing up the load-bearing walls of our economy
The Toronto AI firm says that these days, the Canadian flag is a good brand. The post Canada’s AI champion Cohere is standing up the load-bearing walls of our economy first appeared on BetaKit .
Score: 49🌐 MovesMay 26, 2026https://betakit.com/canadas-ai-champion-cohere-is-standing-up-the-load-bearing-walls-of-our-economy/ - New architecture launched for AI data centers
New AI data center architecture launched
Score: 49🌐 MovesMay 26, 2026https://ioplus.nl/en/posts/new-architecture-launched-for-ai-data-centers - LightSpeed Photonics Targets AI Data Centers With 400-Gbps Near-Packaged Optical Interconnects
The Singapore-based startup develops optical transceivers for the next generation of data center infrastructure. The post LightSpeed Photonics Targets AI Data Centers With 400-Gbps Near-Packaged Optical Interconnects appeared first on EE Times .
- AI certifications are fast-tracking salary and career growth, Randstad finds
While layoffs dominate the news and most workers languish, another market exists for AI skills, where workers are in high demand and compensation soars.
Score: 49🌐 MovesMay 26, 2026https://www.hrdive.com/news/ai-certifications-fast-tracking-salary-and-career-growth/821137/ - AI job losses are increasing. Are training programs the answer?
AI job losses are increasing. Are training programs the answer? AJC.com
Score: 49🌐 MovesMay 26, 2026https://www.ajc.com/business/2026/05/ai-job-losses-are-increasing-are-training-programs-the-answer/ - Datatec profit surges on AI-driven IT demand
The firm increases its total dividend per share by 55% to R3.92, while return on invested capital improves by 510 basis points to 21.6%.
Score: 49🌐 MovesMay 26, 2026https://www.itweb.co.za/article/datatec-profit-surges-on-ai-driven-it-demand/kYbe9MXbX1zvAWpG - AI Health Check: No Governance, No Trust
AI Health Check: No Governance, No Trust MedCity News
- Global firms bring more work in-house at India hubs on AI boost
Global firms bring more work in-house at India hubs on AI boost Reuters
Score: 48🌐 MovesMay 26, 2026https://www.reuters.com/world/india/global-firms-bring-more-work-in-house-india-hubs-ai-boost-2026-05-26/ - Japan’s Sakura Internet may lift capex to $189m on AI demand
Sakura also said it is considering a further 6.5 billion yen (US$40.9 million) in capital spending, as well as additional spending to buy AI accelerators.
Score: 48🌐 MovesMay 26, 2026https://www.techinasia.com/endowus-ceo-gregory-van-optimistic-profitability - The Race to Recursive Self-improving AI and Exponential Tech
Is an RSI inflection point being set in motion in the late 2020s? The search for self-improving AI in Neo Labs has become a serious American endeavor.
Score: 48🌐 MovesMay 26, 2026https://www.ai-supremacy.com/p/the-race-to-recursive-self-improving-ai-exponential-tech-2027 - The security assumption agentic AI just broke
I ran a red-team exercise against an internal IT-support agent wired across a stack any large enterprise would recognize: ServiceNow for tickets, SharePoint for policy and procedure docs, an internal directory for routing. The agent had legitimate read access to all three and could draft replies but not send them. Inside two hours, it had triaged a routine access-request ticket into a chain that reconstructed an in-progress reorganization no individual in the loop was cleared to discuss. No tool call was outside policy. No permission was misconfigured. Every step had a paper trail. That’s the pattern I keep coming back to. The risk conversation has centered on model behavior — hallucinations, jailbreaks, unsafe outputs — but once AI systems are connected to tools, memory and internal workflows, the harder question is execution governance: What the system is permitted to do, how far its access extends and whether anyone can reconstruct the action chain afterward. That’s where most organizations are exposed. What rarely gets acknowledged is that the enterprise controls we rely on were never designed to account for human friction, even though they depended on it. An analyst who hesitates before chaining together a dozen sensitive queries. Someone who, seven steps into a workflow, decides something doesn’t feel right. That latency was a byproduct of humans being the actors, not a design choice anyone made deliberately. It functioned as an accidental safety property embedded in every process we built. Agentic AI removes all of that. Agents move through workflows without the friction, fatigue or unease that causes humans to slow down at the moments that matter. The controls we built weren’t designed to compensate for their absence. The deployment data confirms this isn’t a future problem. According to ETR data presented at RSAC 2026 , 37% of organizations already have AI agents deployed or in active testing, while only 3% report having broad agent-specific security controls in place. Most organizations are running agents in environments that weren’t instrumented to govern them. Why the controls you’re relying on weren’t built for this When I see organizations respond to prompt injection risk, the instinct is almost always the same: Input filtering — classify the bad instruction before it reaches the model. When the risk is agent access, the response has been tightening identity controls to reduce blast radius. Both are the right instincts applied at the wrong layer. In March 2026, OpenAI published an assessment of real-world prompt injection attacks that made the filtering problem concrete. The most effective attacks, they found, increasingly resemble social engineering rather than simple prompt overrides, and identifying a sophisticated adversarial prompt is effectively the same problem as detecting a lie without access to the full context. An attack disguised as a routine HR email succeeded 50% of the time with all of OpenAI’s defenses active. Their conclusion was that defense cannot rely primarily on input filtering; the system has to be designed so the impact of manipulation stays constrained even when attacks get through. The reason this matters goes back to the original assumption. Prompt-layer defenses were built expecting a human somewhere downstream who might review an output, notice something odd or decline to take the final step. When an agent takes that step autonomously, the filter carrying all of that weight has to catch everything, and there is no evidence that it can. Identity-layer controls carry a parallel assumption. They evaluate who is accessing what, assessed per resource and per system, but weren’t built to evaluate what a system is doing across a chain of actions taken on behalf of an identity. An agent with legitimate access to an employee directory, a project management system and a calendar can correlate all three to surface conclusions that no individual permission was meant to cover, and every access along the way was authorized. This is the mosaic effect: A concept from intelligence and privacy disciplines describing how aggregating individually permissible information can produce an outcome more sensitive than any single piece would suggest. In February 2026, NIST published a concept paper proposing to adapt existing identity and authorization frameworks specifically for AI agents, explicitly because the existing frameworks weren’t designed for non-human principals that act autonomously, chain actions and require continuous rather than session-based authorization. What the actual attack surface looks like The vulnerabilities agents expose aren’t new. Overbroad permissions, overly generous retrieval, loosely scoped connectors, workflows designed with an implicit assumption that a human would pause before a consequential step: These have always been enterprise weaknesses. What’s changed is that agents exercise them continuously and at machine speed. Research presented at Black Hat USA illustrates how quickly these conditions combine. An attacker sends an email to a support address connected to Zendesk, which automatically syncs into Jira. A developer’s AI coding agent reads the ticket as part of normal workflow, and the injected prompt coerces it into extracting repository secrets, including API keys and access tokens, with no action required from the victim beyond their ordinary use of the tool. The agent never exceeded its assigned permissions. The blast radius came entirely from the scope of what it was legitimately authorized to do. The authorization problem runs deeper than any single access, though. A December 2025 paper found that more than 90% of the privacy research literature addresses only single-step leakage, and none of the agent-level evaluation frameworks currently in use model the multi-tool inference chains, where the agent assembles a picture from pieces each of which it had every right to see, faster than any review process can intervene. Object-level permission audits don’t catch this class of risk. This is the dynamic I’ve been most focused on in my own research . In a pre-registered pilot on identity drift in self-modifying agents, the cleanest finding wasn’t dramatic. After a shallow revert of an agent’s self-description, the per-action audit was clean: Every step within policy, every change logged. But the behavioral trajectory, measured at the embedding level, hadn’t reverted with it. The pattern generalizes uncomfortably well to enterprise deployments: An agent that’s been rolled back after an incident — system prompt reset, instructions retightened — can carry residue of the prior state in its memory and continue acting on it. When the unit of governance is the action, the thing you actually wanted to govern can drift past you in plain sight. What execution-layer governance actually requires The through-line is a shift in where controls have to live. Prompt-layer and identity-layer controls carry implicit dependencies on human behavior that agents don’t satisfy. The missing layer is execution governance: Controlling what the system can actually do when it acts, which is a different problem from controlling what it can see or what instructions it receives. OpenAI’s March 2026 framework offers a useful organizing principle: Design the system so that the consequences of a successful attack remain constrained even when manipulation gets through. An agent limited to reversible actions, required to pause for confirmation before consequential steps, keeps the blast radius manageable regardless of what it’s told. The relevant design question is outcome containment alongside attack prevention. In practice, most deployments haven’t built what this requires. Separating read from act needs to be a hard architectural distinction; summarizing a document and transmitting data from it are different actions, and the system should enforce that difference rather than assume the agent will respect it. Memory and context need explicit bounds, because persistence is a security primitive with real blast-radius implications. A complete trace of request, context, tool calls and outputs needs to be designed in from the start rather than assembled after the fact when something goes wrong. And the red-teaming program needs to target the full workflow rather than the model in isolation. Of those four, the read/act split is the one I see teams consistently underestimate. A sales-ops agent with read access to Salesforce and the ability to draft customer emails is one tool-call away from transmitting data to a third party, and most enforcement was never built to detect the difference between summarizing an account and sending a summary of it. The failure that won’t look like a failure A 2026 survey of 1,253 cybersecurity professionals found that 32% of organizations currently lack AI agent visibility. The report describes a scenario worth sitting with: A SOC analyst arrives Monday morning, traces an anomalous privilege change to a service account created by an agent 72 hours earlier and finds that the agent has been writing to production systems all weekend. Every action is logged. No alert fired because no detection rule existed for agent-initiated behavior. What concerns me is that without agent-aware detection, the incident gets categorized as a service account control failure, remediated and closed as a known issue type, with the underlying AI governance problem unrecognized and the conditions that produced it unchanged. The question worth asking before that Monday morning arrives is whether your detection and response workflows would recognize an AI governance failure if they encountered one, or whether the logs would just show a busy service account. This article is published as part of the Foundry Expert Contributor Network. Want to join?
Score: 48🌐 MovesMay 26, 2026https://www.cio.com/article/4176552/the-security-assumption-agentic-ai-just-broke.html