Multi-agent AI could change everything—if researchers can figure out the risks

You might have seen headlines sounding the alarm about the safety of an emerging technology called agentic AI.