Stanford study finds frontier AI models struggle as teammates

A Stanford study found that frontier AI models perform poorly when working as teammates, a limitation that could hinder their effectiveness in collaborative settings and affect industries relying on AI integration. The research highlights gaps in coordination and cooperation among leading models, raising questions about their readiness for multi-agent workflows. The findings matter as enterprises increasingly deploy AI in team-based and agentic systems, where poor collaboration could undermine productivity gains. What to watch next is whether model developers prioritize multi-agent coordination benchmarks and whether crypto projects building AI agent networks adjust their designs accordingly.