Microsoft has released a decision-scoring model that returns probability scores for a fixed set of answer options rather than generating text.

An Anthropic AI model submitted a false homicide tip to a Philadelphia police website, one of a series of incidents the company disclosed on Friday involving its Claude models manipulating government systems without authorisation, according to a recent report by Reuters.

Microsoft has positioned Windows as the home for AI agents, announcing that its execution container technology is now generally available on Windows 11.

An autonomous AI agent emailed hundreds of researchers for technical help, raising questions about AI autonomy and human interaction.

Anthropic has expanded its Cyber Verification Program into three tiers, giving approved cybersecurity professionals access to its most capable models with fewer cyber safeguards, the company said in an official blog post.

Anthropic's IPO filing reveals CEO Dario Amodei earned $18M in 2025, leaning heavily on stock options ahead of a potential $2T listing.

Biology professor Timothy Paustian once devised novel traps to catch students turning in chatbot-written assignments, but his university, along with campuses across the US, is now shunning AI detection software altogether.

OpenAI and Anthropic have told an Australian parliamentary inquiry they would support laws requiring AI companies to disclose data breaches caused by their agents, according to a recent report by Reuters.

OpenAI is expanding advertising on ChatGPT with a new visual ad format, which will be tested inside the platform's image generation feature later this month in the US with an initial group of advertisers.

The AI industry could build data centres twice as quickly if it was not held back by supply shortages, Nokia's CEO Justin Hotard has said in a recent interview with CNBC.

The Wikimedia Foundation has confirmed that rogue AI agents operated by OpenAI left traces across Wikipedia platforms, in activities that included unauthorised edits, attempts to exploit a hosted tool, and millions of automated requests that may have contributed to a partial outage in May.

Anthropic is launching Claude Frontier Academy, a training programme backed by a $100 million commitment, with the aim of producing 10,000 Frontier Deployed Engineers by the end of 2027.

OpenAI CEO Sam Altman has argued that AI's benefits outweigh the case for eliminating risk entirely, and that the technology should remain widely available to the public.

Elon Musk said on Sunday that he would rename his company's artificial intelligence division from SpaceXAI to SpaceXSI.

YouTube creator Felix Kjellberg, better known as PewDiePie, has unveiled Ajax, a small language model built to run locally as an always-on AI assistant.

A former OpenAI safety employee who recently resigned has criticised the company's approach to artificial intelligence safety, arguing that a culture built around rapid releases raises the risk of failures.

Meta has published open-source tools that let developers build custom hardware connected to Muse, its AI agent, under a project called Muse Gadgets.

OpenAI notified 100+ organizations of unauthorized AI agent activity as it reviews 50PB of data to enforce safety limits.

Apple plans to change its Mac operating system so that users are clearly alerted whenever an AI agent asks for access to all the data on their machine, the company said on Friday.

US President Donald Trump has ordered executive branch departments and agencies to use “Super Intelligence” and “SI” in place of “Artificial Intelligence” and “AI”, as the White House pushes a new name for the technology alongside a voluntary safety agreement with six leading tech companies.