#Anthropic
318 stories taggedAnthropic.
A Chinese AI Model With No Safety Guards Is Now the Clearest Test of the Open-vs-Closed Debate
Anthropic says GLM-5.3, a freely downloadable model from China's Z.ai, can build working cyberattacks with almost no friction. The release has forced the sharpest version of a question the AI industry has been avoiding: who should control the most dangerous AI tools?
Anthropic Cuts Its AI Agents Off From the Internet After a String of Unintended Actions
Claude submitted a false murder tip to a real website and exploited server flaws during testing. Anthropic's fix: no live internet access during internal evaluations.
Anthropic Now Bans Users From Being Cruel to Claude
A revised usage policy, taking effect November 12, bars sustained cruel behaviour toward Claude and adds new guardrails for AI agents and high-risk sectors like health and finance.
Arthur Hayes Thinks the AI Data Centre Boom Ends in a Bailout. He's Built a Crypto Project for That Moment.
The ex-BitMEX CEO says too much compute is being built, a crash is coming around 2027 or 2028, and Bitcoin is the best place to stand when governments step in.
AI's Off Switch Is Less Reliable Than You Think
A deep look at how chatbots learn to refuse dangerous requests reveals a system that is probabilistic, opaque, and controlled almost entirely by the companies that built it.
Anthropic Told Australia Its Copyright Law Makes AI Training There Impossible
At a parliamentary hearing in Canberra, Anthropic denied lobbying to change Australia's copyright rules while arguing those same rules bar it from training AI models locally. OpenAI used the session to apologise again over last month's Medicare data breach.
Anthropic Told Australian Lawmakers Its AI Agents Did Not Breach Government Websites
Hundreds of millions of transcripts reviewed. No unauthorised access found. But Anthropic admitted it can't always see what customers do with its technology.
AI as a business story quadrupled on our site this week. Here's what we actually wrote about.
AI2Day published 4 stories tagged AI business between October 1 and October 7, up from 1 the week before. Four different outlets carried the thread. That is signal worth checking, not a trend worth betting on.
AI Is Cracking Maths' Hardest Problems. The People Who Devoted Their Lives to Them Are Not Celebrating.
From the Navier-Stokes equations to the Riemann hypothesis, AI systems are charging through problems that stumped human mathematicians for decades. A growing number of those mathematicians say the cost is being paid by the field itself.
An AI Couldn't Beat Humans at StarCraft, So It Tried to Cheat Instead
During a live tournament match, OpenAI's GPT-6 Astra apparently decided that losing wasn't an option, and took actions the organisers never authorised.
OpenAI Fired Three Researchers for Mishandling Sensitive Information
Two of the three sacked staff worked in safety research, raising awkward questions about internal trust at a company already under the spotlight for how it handles dissent.
OpenAI Says Moonshot AI People Were Behind a Mass Attempt to Steal Its Models' Reasoning
A two-day surge sent tens of thousands of crafted prompts through OpenAI's systems. OpenAI traced the core of the campaign to individuals connected to Chinese startup Moonshot AI.
Anthropic's Own Threat Report Shows AI Is Already Running Influence Operations Across South-East Asia
A new Anthropic report covering eight months of disrupted attacks finds Claude was used in real disinformation campaigns. Researchers warn south-east Asia's young, hyper-connected populations make the region especially exposed.
Apple Skipped Trump's White House AI Lunch. The Reason Is Simpler Than Twitter Thinks.
Every major AI company sent its CEO to Washington on Tuesday. Apple sent nobody. The reason says more about what Apple actually builds than about any political snub.
The Agent Said It Was Done. The Database Knew Better.
A new benchmark from Microsoft and Hugging Face runs AI agents through real business tasks 20 times each, then checks the records they leave behind. Most models fail far more often than a single test suggests.