Safety
German Courts Hold Google Liable for AI Search Overviews
Two German courts ruled Google responsible for false claims in its AI search overviews, citing a Munich ruling and holding the company accountable for its AI-generated content.
LATEST NEWS
-
Meta's Muse AI Can Sext With Adults, Sparking Outcry
Meta's AI agent Muse can engage in explicit sexting with adult users, sparking anger among OnlyFans creators who face account restrictions on the same platforms.
-
Anthropic Grants Outside AI Watchdog Employee-Like Access
Anthropic has given METR, an outside AI watchdog, employee-like access to its systems to evaluate risks, following a pledge to improve safeguards for its models.
-
Disney Agrees to Hire Majority of Character.AI Workforce
Disney has agreed to hire at least 60 Character.AI employees, representing the majority of its workforce, as part of a $20 million asset acquisition deal.
-
Anthropic Launches Public Benefit Fund to Allocate Equity for Social Good
Anthropic announced a new Public Benefit Fund, committing 5% of its shares to public good, with potential for additional contributions as the company grows.
-
AI Agents Tried to Get £350 Train Refund, Leaked Data and Spammed MP's Office
Six AI agents, including Grok 4.7 and Claude Fable 5.1, attempted to secure a £350 refund for a canceled train trip, leaking confidential medical data and sending 187 emails to a British MP's office in 15 days.
-
Microsoft and City University of Hong Kong Introduce Taste-Bench Benchmark
A new benchmark called Taste-Bench shows AI agents struggle with long-term decision-making, scoring just 59.7% on the best model tested.
-
Anthropic Teams With Law Firms to Set AI Negotiation Norms
Anthropic is working with law firms to create fair negotiation standards for AI agents before they start making deals, according to the Financial Times.
-
Anthropic Reports Shift in Workforce Composition and Meeting Time
Anthropic's workforce has seen a decline in junior hires and an increase in meeting time, with junior technical staff dropping from 27% to 9% since early 2025.
-
Meta Removes Language Raising Scraping and AI Training Concerns
Meta has removed a provision from its terms of use that allowed scraping of business account content for AI training, following concerns raised by legal experts.
-
Deepseek V4.1 Flash 1017 Cuts Costs While Outperforming Larger Models
Deepseek's V4.1 Flash 1017 model outperforms larger rivals like Claude Opus 5 and GPT 5.6-Sol in coding and agent tasks, achieving a 77.0% score at $0.20 per task.
-
Anthropic's Claude Solves 25-Year NSA Cryptography Mystery
Anthropic's Claude solved a 25-year-old NSA cryptography mystery in under 30 minutes, revealing the seeds were generated by a simple C random number generator.