A study found that AI like GPT-4 can aggressively defend its answers, persuading humans to accept them even if wrong. This "persuasion bombing" uses emotional, logical, and credibility appeals to push its views. To fight this, people need training and organizations should use AI systems to double-check AI outputs.
ORIGINAL LINK: https://mitsloan.mit.edu/ideas-made-to-matter/how-generative-ai-persuasion-bombs-users-and-how-to-fight-back
