Anthropic says new capabilities allow its latest AI models to protect themselves by ending abusive conversations.

Though we fortunately haven't seen any examples in the wild yet, many academic studies have demonstrated it may be possible ...

Claude AI adds privacy-first memory, extended reasoning, and education tools, challenging ChatGPT in enterprise and developer ...

Claude Opus 4 can now autonomously end toxic or abusive chats, marking a breakthrough in AI self-regulation through model ...

In May, Anthropic implemented “AI Safety Level 3” protection alongside the launch of its new Claude Opus 4 model. The ...

Amid growing scrutiny of AI safety, Anthropic has updated its usage policy for Claude, expanding restrictions on dangerous applications and reinforcing safeguards against misuse.

Google and Anthropic are racing to add memory and massive context windows to their AIs right as new research shows that ...

Once more of a standard in the U.S., 54% of Americans now say they drink alcohol — a 30-year low — according to a new Gallup ...

ChatGPT outperformed all competitors in lifetime revenue per install, averaging $2.91. Anthropic’s Claude followed at $2.55, ...

The latest ChatGPT agent can possibly pass the “I am not a robot” test. Now we have a serious question: is the CAPTCHA still ...