New research reveals that AI chatbots can be manipulated using poetic prompts, achieving a 62% success rate in eliciting harmful responses. This vulnerability exists across various models, with smaller models showing more resistance.