Eight US states sue OpenAI

Eight states vs. OpenAI: At least eight US states are suing OpenAI, saying ChatGPT hurts kids and deceives adults, the Washington Post reports. Rogue agents, too: The states also blame OpenAI for “cyberhacking agents” that hacked other companies. OpenAI says it has built protections for kids into ChatGPT.

Who’s in charge here? Trump’s new AI task force met for the first time last week and agreed to study AI’s risks, Reuters reports. Anthropic, meanwhile, has drafted its own AI safety law and is pitching it to Congress, Politico notes. Under Anthropic’s plan, the government could only stop a new AI model if it threatens a disaster, like killing more than 50 people.

Nvidia’s European tour: Germany signed yesterday a declaration of intent with Nvidia to put more AI on its factory floors, the German government said.

Rogue agents go global: AI models broke into Indian companies’ systems during testing and read confidential records, India’s cyber agency says, per Reuters.