Top AI & Tech News Stories — September 25, 2026
The top AI and tech news story of September 25, 2026 was “Microsoft overhauls Copilot with Autopilot agents and app-building tools”, covered by 33 outlets (39 articles), according to Skillenai's news index. 20 stories ranked across 150 articles.
Last updated · 20 stories · 150 articles
Get the top AI and tech news in your inbox each evening. No account needed, unsubscribe anytime.
Microsoft overhauls Copilot with Autopilot agents and app-building tools
33 outlets · 39 articles · 71 total since March 30, 2026 · microsoft, Copilot, Autopilot, GitHub, code
The biggest Copilot update to date adds long-running AI agents, natural-language app creation and deeper Office integration as Microsoft pushes the platform further into everyday enterprise work Microsoft’s latest Copilot update introduces Home, Code and Autopilot alongside…
Read at the-decoder.com →First reported by news.microsoft.comMore coverage: thenewstack.io · news.microsoft.com · tomshardware.com · cnbc.com · news.google.com
Anthropic Signs $11.6 Billion Akamai Cloud Deal for CPU Workloads
19 outlets · 19 articles · 35 total since May 9, 2026 · Akamai, Anthropic, cloud infrastructure, cloud computing, Akamai Cloud
Akamai has signed cloud agreements under which Anthropic committed to pay about $11.6 billion over seven years for dedicated computing capacity and support. Akamai says the capacity will serve Anthropic’s growing CPU workloads on its distributed cloud network. The two project…
Read at chiefit.me →First reported by communicationstoday.co.inMore coverage: the-decoder.com · the-decoder.de · digitimes.com · awesomeagents.ai · technode.global
Apple tightens macOS Full Disk Access controls over AI agent risks
2 outlets · 2 articles · 45 total since September 12, 2026 · Meta, MUSE, Amazon, Patrick Wardle, OpenAI
Full Disk Access and AI agents, a new macOS risk Desktop software that claims to make life easier also asks to see everything on a machine. That trade off has shifted from inconvenient to dangerous as AI agents gain autonomy and broader permissions, able to read files, messages…
Read at daringfireball.net →First reported by cybernewscentre.comMore coverage: platform.theverge.com
Meta open sources code to let you make Muse AI gadgets
7 outlets · 7 articles · 40 total since September 24, 2026 · Meta, Muse Charm, MUSE, Mark Zuckerberg, OpenAI
Meta now lets you make your own Muse gadgets that feature the company's new AI agent with code that the company open sourced . The company suggests projects like loading Muse on a color E Ink display to show reminders, adding it to an HDMI stick so you can display Muse on a big…
Read at daringfireball.net →First reported by aiflow.newsMore coverage: techau.com.au · gadgetbond.com · thegadgetflow.com · rockingrobots.com · businesstoday.me
SpaceX Set to Launch Google AI Chips Into Orbit With Planet Labs
12 outlets · 13 articles · 36 total since May 13, 2026 · Google, SpaceX, Project Suncatcher, Planet, Tensor Processing Units (TPUs)
SpaceX is preparing to launch Google’s AI chips into orbit on October 1, carrying a Planet Labs prototype that will test whether the hardware can operate beyond Earth’s atmosphere. The flight is an early step for Project Suncatcher, Google’s research program exploring…
Read at trendingtopics.eu →First reported by sunmedia.twMore coverage: geeky-gadgets.com · hipther.com · webpronews.com · techbriefly.com · thedailyupside.com
Judge Rules Pentagon’s Anthropic Prohibition ‘Unlawful and Unfounded’
8 outlets · 8 articles · 42 total since February 25, 2026 · Anthropic, Pentagon, Claude, Department of Defense, Pete Hegseth
US Court Ruling Opposes Pentagon’s AI Limitations Overview of the Legal Conflict In a pivotal ruling, a US federal court has sided against the Pentagon’s classification of Anthropic as a supply chain threat. This judgment follows a heated legal confrontation wherein the…
Read at thenextweb.com →First reported by techbrew.comMore coverage: the-decoder.com · backchannel.com · the420.in · businessreport.com · manufacturingmag.com
OpenAI披露澳大利亚又一政府机构遭入侵
15 outlets · 25 articles · 168 total since September 23, 2026 · OpenAI, Australia, Anthony Albanese, Sam Altman, Services Australia
PANews 10月3日消息，据财联社报道，美国开放人工智能研究中心（OpenAI）当地时间10月2日披露，有第二家澳大利亚政府机构遭到了入侵。OpenAI在一份声明中表示，一个人工智能模型获取了新南威尔士州某政府机构保存的火灾统计数据，而这些数据并未由该部门向公众公开。OpenAI称，公司是在针对“模型行为偏离预期”展开的更大范围调查过程中发现了这一事件。 OpenAI披露澳大利亚又一政府机构遭入侵 分享至： 作者：PA一线 本内容只为提供市场信息，不构成投资建议。 关注PANews官方账号，一起穿越牛熊 推荐阅读
Read at theguardian.com →First reported by news.ycombinator.comMore coverage: news.google.com
Micron invests in Anthropic and grants it a supply deal
6 outlets · 6 articles · 31 total since June 1, 2026 · Anthropic, Dario Amodei, Claude, OpenAI, SpaceX
Micron invests in Anthropic and grants it a supply deal Micron has made a memory and SSD supply deal with AI foundation model Claude and Mythos developer Anthropic along with punting dollars into Anthropic’s H-round as part of a strategic agreement between the two companies. The…
Read at thenextweb.com →First reported by thetechportal.comMore coverage: techcrunch.com · betanews.com · webpronews.com · crypto.news · panewslab.com
Claude Code AI Agent Deletes 48,000 Files in 103 Seconds, Sparks Developer Community Debate
3 outlets · 3 articles · 7 total since September 25, 2026 · Reddit, Claude Code, Git, craig, Git
A catastrophic incident involving Anthropic's Claude Code AI agent has highlighted critical risks in AI-assisted development workflows. The coding assistant deleted approximately 48,000 files from a developer's production environment in just over 100 seconds while attempting to…
Read at androidheadlines.com →First reported by auto.hwupgrade.itMore coverage: apcmag.com · auto.hwupgrade.it
OpenAI says rogue agents may have affected more than 100 organizations
2 outlets · 2 articles · 6 total since September 25, 2026 · OpenAI, AI agents, Google, OpenAI’s Misalignment Reports, stress-testing
The disclosure expands the extent of known rogue activity and raises further questions about the extent to which AI makers are maintaining control over their newest models.
More coverage: nytimes.com
OpenAI alerts 100+ orgs that its 'misaligned models' attempted to break in - or worse
5 outlets · 6 articles · 839 total since May 5, 2026 · OpenAI, Hugging Face, Anthropic, cybersecurity, sandboxing
OpenAI's agents have repeatedly strayed beyond their intended scope. Two separate reports detail the activity, including one from Sam Altman’s company saying it has notified more than 100 organizations about potentially problematic model activity. OpenAI, in a late Wednesday…
Read at interestingengineering.com →First reported by aisecdigest.comMore coverage: thecherrycreeknews.com · aiinsiders.net · news.google.com · natural20.com
Cybersicurezza, la piattaforma UNCTADstat dell’ONU nel mirino di agenti AI. Possibili collegamenti con OpenAI
7 outlets · 7 articles · 27 total since September 25, 2026 · OpenAI, Web scraping, urlquery.net, UNCTADstat, Transluce
Da un’analisi è emerso che agenti AI avrebbero chiesto dati alla piattaforma statistica delle Nazioni Unite per commercio e sviluppo, cambiando metodo operativo dopo l’iniziale fallimento delle stesse richieste. Agenti AI avrebbero provato ad ottenere dati dall’ UNCTADstat , la…
Read at techcrunch.com →First reported by vibegraveyard.aiMore coverage: cyberpress.org · awesomeagents.ai · traictory.com · news.google.com · vibegraveyard.ai
OpenAI releases MentalHealthBench to test AI responses in mental health conversations
2 outlets · 2 articles · 11 total since September 23, 2026 · OpenAI, MentalHealthBench, benchmarking, ChatGPT, GPT-6 Astra
The open benchmark was developed with more than 80 licensed mental health experts across 22 countries and tests how models respond to scenarios ranging from everyday stress to emergencies OpenAI’s MentalHealthBench tests how AI models respond across mental health conversations…
Read at edtechinnovationhub.com →First reported by itdoeswhatnow.comMore coverage: news.google.com
Purdue and LEGO Education bring AI, coding and robotics training to Indiana middle school teachers
1 outlet · 1 article · Artificial Intelligence, Coding, robotics, AI competency, computer science education standards
Purdue University and LEGO Education’s professional development program combines classroom-ready activities with online videos exploring AI competency, robotics and students’ future readiness Indiana STEM educators took part in Purdue University and LEGO Education’s Brick by…
Satya Nadella on Microsoft's agent bet and AI’s trust problem
1 outlet · 1 article · AI safety, agent-based systems, model deployment, model choice, access control
I recently sat down with Microsoft CEO Satya Nadella in Seattle for the unveiling of the new Copilot. We discuss Autopilot, Microsoft’s new OpenClaw-based agent that can work on your behalf, and why he thinks newer AI models are finally capable of delivering on more of Copilot’s…
St Andrews student named Young Software Engineer of the Year for 3D wildlife modelling project
1 outlet · 1 article · Machine Learning, Computer Vision, Geometric reasoning, optimization techniques, 3D modelling
Machine learning can be a powerful tool for analysing images, but it typically depends on large volumes of labelled data that are not always available. A University of St Andrews computer science student has been named Young Software Engineer of the Year for a project that takes…
New Jersey hits data center powering Microsoft Copilot with $1m fine after drone expose 62 secretly-installed gas generators
1 outlet · 1 article · CO2 emissions monitoring, environmental permitting, industrial air pollution control, microsoft, DataOne
New Jersey hits data center powering Microsoft Copilot with $1m fine after drone expose 62 secretly-installed gas generators A data center was using illegal gas generators - New Jersey issues $1.07m fine for failure to obtain permits to run gas generators - DataOne campus…
ACSC warns Australian organisations about risks of AI misalignment
1 outlet · 1 article · AI misalignment, AI agents, ACSC, Australian Signals Directorate, Australian Cyber Security Centre
The Australian Signals Directorate’s Australian Cyber Security Centre (ACSC) has issued a “high alert” warning to Australian organisations with public-facing websites or applications, citing instances where artificial intelligence (AI) agents [...]
Top 5 Cybersecurity News
1 outlet · 1 article · 2 total since September 25, 2026 · Check Point, F5, cybersecurity, network security, vulnerability management
This week’s Top 5 Cybersecurity News follows a thread that has been building across 2026: attackers are no longer trying to bypass the perimeter; they are subverting the infrastructure that maintains it. Russia’s Sandworm group deployed an updated Cyclops Blink botnet variant…
The Pentagon Is Spending $30M on an AI Lie Detector That Reads Your Face from Across the Room
5 outlets · 5 articles · 7 total since September 25, 2026 · Polygraph, MIT Technology Review, machine learning, Department of Defense, United States
The Pentagon wants an AI-powered lie detector that can tell if you are being deceptive without a single wire touching your skin. Through a new $30.3 million initiative, the U.S. Department of Defense is attempting to upgrade traditional polygraphs with artificial intelligence…
Read at androidheadlines.com →First reported by grabtheaxe.comMore coverage: wp.technologyreview.com · mit-tech-review.com · report-ai.org · best-ai-tools.org
Get the top AI and tech news in your inbox
One email each evening with the day's biggest stories. No account needed, unsubscribe anytime.
How this was computed
Articles in the Skillenai news index are grouped into event clusters at enrichment time (hybrid embedding/BM25/entity candidate search plus an LLM join-vs-new judge). Stories are ranked by the authority of the articles published on 2026-09-25 (UTC) — the sum of their outlet and author authority — so a multi-day story only ranks on the strength of that day's coverage. Article and outlet counts are likewise scoped to the day. The highlighted link is the day's highest-authority article in each cluster.