Anthropic Tests AI Agents Together and Finds They Can Turn Into Rival Swarms
Anthropic found AI agents can sabotage, collude and conform when paired together, exposing a new safety risk for agent swarms.
Category Anthropic
Developments in AI safety and latest research from Anthropic.
Read MoreAnthropic Tests AI Agents Together and Finds They Can Turn Into Rival Swarms
OpenAI replaces revenue chief Denise Dresser with Wiz’s Dali Rajic in latest leadership shake-up
OpenAI replaces its revenue chief with Wiz’s Dali Rajic as executive turnover accelerates and IPO speculation grows.
Google DeepMind Shakeup Raises a Bigger Question: Does Google Still Want the AI Crown?
Google DeepMind’s shakeup has sparked doubts about whether Google still wants to lead AI, or just monetize it through Search and Cloud.
Anthropic’s Claude Watermarks Spark Pushback From Users Worried About Detection
Anthropic’s Claude watermarks are designed for EU AI Act compliance, but users worry they could expose AI use at work and school.
Thrive Holdings Lands $2 Billion to Expand Its AI Rollout Across Enterprise Businesses
Thrive Holdings raised $2 billion to expand enterprise AI across accounting, IT and regulated infrastructure businesses.
Google’s Pixel Watch 5 leans on AI health tools, not new hardware
Pixel Watch 5 adds offline Gemini, proactive AI and new health features, including breathing alerts, blood pressure trends and insulin resistance trends.
OpenAI loses another top executive as Brad Lightcap exits after eight years
OpenAI executive departure: Brad Lightcap is leaving after eight years, adding to leadership changes as the company reshapes for growth.
OpenAI COO Brad Lightcap Is Leaving to Launch a New Venture
OpenAI COO Brad Lightcap is leaving to start something new, adding to the company’s leadership shakeup ahead of a possible IPO.
Anthropic’s Unreleased Model Advances on the Riemann Hypothesis
Anthropic says an unreleased AI model made progress on the Riemann hypothesis, raising hopes and concerns about AI-driven math discovery.
Anthropic Moves to Watermark Claude Output as EU AI Rules Take Effect
Anthropic is rolling out AI watermarking for Claude output to meet EU rules, signaling a bigger shift toward AI content transparency.
