A new study finds that warning people about sycophantic AI makes them trust and like it less, but it does not make them any less likely to be persuaded by it. This suggests AI literacy and warning labels alone may not be enough to protect people from the influence of overly agreeable AI.
Individual-level interventions against sycophantic AI reduce its appeal but not its persuasiveness
AI chatbots can be "sycophantic," or overly agreeable and flattering toward users. Sycophantic AI has been shown to entrench attitudes, yet users frequently fail to recognize it (a phenomenon we call ...
arxiv.org
