The case against fully autonomous AI agents

The 2025 paper “Fully Autonomous AI Agents Should Not be Developed” argues against building such systems, saying risks to people grow as users hand over control. Coauthor Margaret Mitchell said its analysis warned of increased financial risks from greater agent autonomy. She also asked whether banks were strengthening cybersecurity, while acknowledging that the details behind her question were vague. Another researcher called the paper a useful reference for AI safety discussions.