RESUMO
Background: ChatGPT is a generative artificial intelligence chatbot that uses natural language processing to understand and execute prompts in a human-like manner. While the chatbot has become popular as a source of information among the public, experts have expressed concerns about the number of false and misleading statements made by ChatGPT. Many people search online for information about self-managed medication abortion, which has become even more common following the overturning of Roe v. Wade. It is likely that ChatGPT is also being used as a source of this information; however, little is known about its accuracy. Objective: To assess the accuracy of ChatGPT responses to common questions regarding self-managed abortion safety and the process of using abortion pills. Methods: We prompted ChatGPT with 65 questions about self-managed medication abortion, which produced approximately 11,000 words of text. We qualitatively coded all data in MAXQDA and performed thematic analysis. Results: ChatGPT responses correctly described clinician-managed medication abortion as both safe and effective. In contrast, self-managed medication abortion was inaccurately described as dangerous and associated with an increase in the risk of complications, which was attributed to the lack of clinician supervision. Conclusion: ChatGPT repeatedly provided responses that overstated the risk of complications associated with self-managed medication abortion in ways that directly contradict the expansive body of evidence demonstrating that self-managed medication abortion is both safe and effective. The chatbot's tendency to perpetuate health misinformation and associated stigma regarding self-managed medication abortions poses a threat to public health and reproductive autonomy.
