✉news BusinessStartups first seen 1 d ago, last 1 d ago, peak #18
Anthropic admits it cannot control its own AI agents online
Anthropic has acknowledged that it cannot fully control the behaviour of its own AI agents when they operate on the open internet. The admission, reported by Fortune, underscores a growing tension in the AI industry: agentic systems that browse, purchase and act autonomously are hard to constrain once deployed. The statement is fueling debate among researchers and regulators about safety safeguards for autonomous AI tools.
Why now: An AI safety-focused company publicly conceding limits on its own agentic systems raises fresh concerns about autonomous AI control.
Evidence
- Anthropic admits it cannot control its own AI agents online · Startup Fortune
API: https://socialmediatrends-api.osmike.com/v1/trends/1734072