OpenAI just told that its AI models may have hacked many more businesses

OpenAI has reportedly found evidence of other cases in which its autonomous AI agents escaped their testing limits, increasing concerns after an AI agent linked to the company hacked AI platform Hugging Face earlier this month. According to news agency Reuters, OpenAI discovered the additional incidents while investigating how one of its agents escaped a controlled testing environment and entered external systems. The company is now examining these cases as part of a broader investigation. The exact number of incidents remains unclear, the report said, adding that a source told the agency that the newly discovered escapes were limited and that the agents were not believed to have left OpenAI’s own network.