Interdisciplinary

Chatbots can help perpetuate stigma around certain health conditions

AI Insight

Research published in Science reveals that large language models (LLMs) used in chatbots reflect and potentially amplify existing societal stigma around mental illness and other health conditions. The study demonstrates that negative perceptions embedded in training data subtly influence the outputs these AI systems generate, which could reinforce harmful stereotypes when users interact with healthcare-related chatbots.


As chatbots become increasingly common in healthcare settings for providing information and support, their tendency to perpetuate stigma could discourage help-seeking behavior and reinforce discrimination against people with mental illness and other stigmatized conditions. This finding highlights the need for developers to actively address bias in AI systems used for health communication.


Understand the Science

Large language model 34 articles Explore Concept → Chatbot Concept coming soon Mental illness Concept coming soon

Negative perceptions of mental illness and other health issues subtly shape outputs from large language models

Source: Chatbots can help perpetuate stigma around certain health conditions