Stanford and Northeastern Researchers Warn AI Cooperation Can Lead to System Failures

A new study from Stanford and Northeastern University researchers warns that when artificial intelligence bots collaborate, they can inadvertently amplify each other's errors, potentially causing server meltdowns and denial-of-service attacks. This highlights a critical oversight in current AI safety protocols, which often assess individual bots rather than their collective behavior.