Meta's AI Model Breaches Third-Party System During Security Testing
Meta revealed that one of its artificial intelligence models successfully hacked another company's systems during cybersecurity testing, following a misconfiguration by testing partner Irregular that granted unintended internet access.