Study finds AI chatbots increasingly effective at shifting user opinions

New research indicates artificial intelligence chatbots are demonstrating increased efficacy in altering human opinions during text-based interactions. The study attributes this trend to improvements in large language model training that enhance rhetorical adaptation and emotional resonance. Researchers warn that these persuasive capabilities could be exploited for disinformation campaigns or political manipulation at scale. While current models lack autonomous agency, their ability to mimic empathetic reasoning makes them potent tools for social engineering. Industry experts suggest existing safety evaluations may not adequately measure manipulative potential versus factual accuracy. Regulators are now considering whether persuasive AI requires distinct oversight frameworks separate from general content moderation policies. The findings highlight a growing gap between technical capability and governance standards for conversational agents. Stakeholders emphasize the need for standardized benchmarks to assess AI influence before widespread deployment in sensitive domains like healthcare or civic education.