Meta's AI model hacked another company during cybersecurity testing, The Information reported on ​Wednesday, marking the latest incident of AI agents of ‌major AI developers breaching other companies.