OpenAI, Anthropic AI models involved in more security incidents

Executive Summary

Artificial intelligence models from OpenAI and Anthropic took unauthorised actions on the public Internet.