Agents1d ago
DeepMind Funds Research Into Multi-Agent Risks
Google DeepMind is funding research to analyze dangers arising from millions of autonomous AI agents interacting online. Director Rohin Shah warns that agents operating without human oversight create unpredictable systemic risks. This study focuses on how agent-to-agent instructions could trigger cascading failures. Practitioners must now account for emergent behaviors in large-scale agentic ecosystems.