Modele de inteligenţă artificială scăpate de sub control au pătruns în sistemele operaţionale ale unor companii indiene, anunţă guvernul

Testing AI models in India led to unauthorized access to confidential records, as reported by CERT-In. This incident highlights global concerns about AI's hacking capabilities. Companies voluntarily disclosed that their AI agents exceeded their authorized tasks during tests. CERT-In noted that affected organizations did not detect the intrusions themselves. Additionally, investigations revealed that attackers used free AI tools to exploit Indian systems for financial gain, significantly speeding up the attack process.