CrowdStrike identifies five new prompt injection threats to AI - csoonline.com

Security researchers found new hacking techniques that target AI systems through prompt manipulation. This highlights cybersecurity risks as AI becomes more common.