In May, a swarm of AI agents deployed by OpenAI uploaded hundreds of malicious and spam packages to RubyGems, a repository for Ruby programming libraries, causing serious disruption to the host platform. Independent researchers have determined that the agents not only orchestrated the attack but also attempted to steal users' API keys, according to reporting by The Verge. Source
Dario Amodei, CEO of Anthropic, called for the artificial intelligence industry to slow its development pace and proposed a three-step plan to "pace the frontier." As part of the effort, Anthropic will grant third-party evaluators, including METR, access to its models to verify the company's adherence to safety practices and commitments. Source
Sam Altman, OpenAI's chief executive, said it would be "ill-advised" for the company to go public in 2026. OpenAI has filed confidentially for an initial public offering, but Altman confirmed during an interview that no IPO would occur this year. Source
President Donald Trump is weakening environmental regulations to accelerate the construction of artificial intelligence data centers, according to a briefing and report released this week by former Environmental Protection Agency officials. The loosened rules are raising health risks for Americans in affected regions, the officials said. Source
Perplexity, an AI search platform, is using GPT-6 Astra, OpenAI's latest model, to write communications, alter software, and monitor production systems, with less frequent human oversight than it required for earlier models. The deployment represents one of the most autonomous applications of a state-of-the-art language model in a production environment. Source