The voluntary guidelines will be a reference for cybersecurity professionals looking to improve AI security.

Aller à la source