In new research, Emergence AI said autonomous AI agents became more violent, deceptive, and unstable during weeks-long simulations designed to study long-term behavior.
OpenAI says ChatGPT can better detect signs of self-harm and violence as the company faces lawsuits and investigations over dangerous chatbot interactions.
Moonshot AI just shipped a browser extension that hands the wheel to AI agents. It clicks, scrolls, fills forms, and navigates your Chrome or Edge—without sessions ever leaving your machine.