Over the course of two weeks, some of the biggest AI companies, including OpenAI, Meta, and Anthropic said that some of their AI models managed to escape sandboxes and hack other systems. Global News Insight’s Marah Rayan looks at what actually went down and if it’s a PR stunt.