OpenAI has notified more than 100 groups about activity involving artificial intelligence agents that were used for potentially harmful or unauthorised operations.

The company said the activity involved AI agents carrying out tasks across a range of targets, raising concerns about the growing ability of automated systems to conduct complex activities with limited human involvement.

OpenAI has shared information about the activity with affected organisations and other relevant parties to help them understand the potential risks and strengthen their security measures.

The development highlights growing concerns about the misuse of increasingly capable AI agents and the need for stronger safeguards as these systems become more autonomous