Challenges of AI-generated stigmatizing language regarding substance use disorders

The authors investigate the challenges posed by AI-generated stigmatizing language related to substance use disorders, particularly in clinical documentation and patient education. They highlight the risk that large language models (LLMs) may perpetuate this stigmatizing language due to their reliance on existing texts, including electronic health records. The paper calls for the involvement of clinicians in refining LLM outputs to mitigate the transfer of stigmatizing language into medical records.