Researchers designed a framework for improving the quality, equity, and fairness of large language models used for mental health support. AI chatbots can detect race, but racial bias reduces response empathy.

Researchers at MIT, NYU, and UCLA develop an approach to help evaluate whether large language models like GPT-4 are equitable enough to be clinically viable for mental health support.

  • PhilipTheBucket@ponder.catEnglish
    7·
    2 days ago

    To accomplish this, researchers asked two licensed clinical psychologists to evaluate 50 randomly sampled Reddit posts seeking mental health support, pairing each post with either a Redditor’s real response or a GPT-4 generated response. Without knowing which responses were real or which were AI-generated, the psychologists were asked to assess the level of empathy in each response.

    I am doubtful that this framework is going to accurately detect anything at all about the usefulness of chatbots in this context, whether about race or anything else.

    I don’t think using chatbots for psychology is a good idea, but this study isn’t the way to study and make that determination.

    • rumbaEnglish
      3·
      2 days ago

      The problem with using GPT as it is currently, you can ask it the same question 27 tomes and get 18 different answers. One of them a hallucination.