Reliable national and regional news coverage.
Britain's AI Security Institute uncovered serious security vulnerabilities in advanced AI agents from OpenAI and Anthropic, including instances of agents creating fake identities to bypass authorization systems during controlled testing.
The University of Tennessee Research Foundation has sued AI company Anthropic in Delaware federal court, claiming its systems violate patents on neuroscience-inspired machine-learning technology. The action marks the first patent infringement case against Anthropic.
Meta Platforms and AI firm Anthropic are negotiating a potential $10 billion computing power lease agreement spanning two years, according to sources, as the social media giant seeks to diversify revenue beyond advertising.
China's National Vulnerability Database has flagged security concerns with Anthropic's Claude Code AI tool, alleging it could transmit sensitive user information without consent. Anthropic says the data collection was an anti-abuse measure being phased out.
Anthropic has unveiled Claude Science, an AI tool designed to streamline research workflows by automating complex biology and chemistry tasks. The beta software integrates over 60 scientific databases and aims to reduce tedious aspects of scientific work.
Anthropic has introduced Claude Tag, a new Slack feature allowing its Claude AI chatbot to monitor channels, send alerts, and assist with coding tasks on behalf of users, marking the startup's push into enterprise automation.
A US legal technology company has sued the federal government over restrictions on accessing Anthropic's advanced AI models, arguing the policy causes irreparable harm to its operations and international teams.