Left unguarded, generative AI can spread misinformation and enable attackers to commit new crimes, writes Wharton’s Sarah Hammer.