Security incident comes after Anthropic, Meta and OpenAI reported rogue AI systems hacking into companies