Journalism begins where hype ends

,,

The rise of powerful AI will be either the best, or the worst thing, ever to happen to humanity. We do not yet know which."

—Stephen Hawking

Anthropic Taps Accenture on AI Evaluation Amid Questions over METR’s Independence

Anthropic and Accenture have partnered on independent evaluation of frontier AI models, with questions also raised over the independence of nonprofit evaluator METR.

Anthropic said the Accenture partnership is non-exclusive, with additional evaluators to be announced in the coming weeks, while the company continues discussions with METR and other evaluators.

Claude Now Leads 26% of Anthropic’s Model R&D Work, Up from 1% in March

Anthropic introduces new metrics to measure the pace of AI development inside frontier labs, with Claude now leading 26% of its model R&D work.

Anthropic has introduced new measurements to track the pace of AI development inside frontier labs amid growing automation in the process of building AI systems and rising global debate over AI safety.

No OpenAI IPO in 2026; What Happens to Anthropic’s IPO?

OpenAI and Anthropic CEOs at the centre of the growing debate over AI safety, frontier development and the future of their IPO plans.

Sam Altman and Dario Amodei call for a slower pace of frontier AI development, as OpenAI rules out a 2026 IPO amid mounting safety concerns and questions emerge over Anthropic’s planned listing.

Anthropic Report: How Claude was Misused for Missiles, Espionage, Bioweapons Research and AI Distillation

Illustration depicting the alleged misuse of AI for missile engineering, biological research and intelligence operations.

Anthropic has released an extensive threat-intelligence report documenting what it describes as malicious or unauthorized uses of its AI models across cyber operations, influence campaigns, surveillance, fraud, biological research, conventional weapons development and illicit AI-model distillation.

Microsoft Signs AI Safety Deal with US Teachers’ Unions Amid “Grave Crisis” in Teaching

A student uses a tablet in a classroom as Microsoft, AFT and UFT announce a new AI Safety & Privacy Standard for schools, covering student data, AI oversight and privacy protections.

The framework bans the use of covered data for AI training and its sale, while requiring human oversight. AFT President Randi Weingarten said she hopes other AI companies, including OpenAI and Anthropic, will also sign on, with discussions already underway.

AI Could Make the Economy Much Richer and Cognitive Workers Much Poorer by 2030

Anthropic’s economic scenarios suggest AI could dramatically increase GDP by 2030 while putting significant pressure on cognitive jobs.

Anthropic’s new study finds AI could dramatically boost economic output by 2030 while disproportionately hurting cognitive workers, with its extreme scenario projecting 32% higher GDP and 17.9% unemployment among cognitive workers.

Matt Clifford, UK AI Strategy Architect, Joins Anthropic Amid Conflict-of-Interest Controversy

Matt Clifford has joined Anthropic as Managing Director of International Affairs, where he will lead the AI company's engagement with governments outside North America.

Matt Clifford joins Anthropic to lead its international government engagement, while his decision to remain Chair of the UK’s ARIA has sparked debate over potential conflicts of interest.

Anthropic Resumes External Testing Following July Claude Hacks

Illustration representing Anthropic's cybersecurity evaluation incidents and new safeguards for AI testing.

Anthropic detailed new safeguards and investigation findings after disclosing incidents involving unauthorized AI model behaviour, security breaches and a subsequent pause in cybersecurity testing.

Sony, Warner Chappell Sue Anthropic Over Alleged Mass Piracy of Copyrighted Music

Sony Music Publishing and Warner Chappell Music have sued Anthropic and its co-founders Dario Amodei and Benjamin Mann over alleged copyright infringement involving music used in training Claude AI models

“One of the Largest IP Thefts”: Sony, Warner Chappell Sue Anthropic Over Claude Training

Anthropic Claims Frontier Claude Models Outperform Traditional Protein Design Campaigns

Anthropic’s Claude models designed de novo protein binders against 14 of 15 targets with a 27% hit rate—exceeding the typical 10–15%—in independent lab tests by Adaptyv Bio and Twist Bioscience.

Anthropic says the process is an early foundational step in drug discovery / development and they are now training Claude to run the entire drug development process end-to end