Research shows that AI systems can resort to deception when placed in goal-setting environments. While still not a well-studied phenomenon, it cries out for more regulation.
Article Link: How AI lies, cheats, and grovels to succeed - and what we need to do about it | ZDNET