AI chatbots spread inaccurate abortion info, prompting activist countermeasures

Chat-based AI systems are increasingly being used for health queries, including abortion care, but they frequently deliver inaccurate or biased information. After the 2022 Supreme Court decision that ended federal abortion protection, misinformation online surged, and the entry of large-language models has compounded the problem. Organizations like the Guttmacher Institute, Plan C Pills and I Need an A see more traffic from AI referrals yet struggle with bots that misrepresent legal options or direct users to crisis pregnancy centers.

Studies show only about two-thirds of ChatGPT-3.5 answers on abortion are acceptable, and some models actively spread falsehoods about medication abortions. Advocates are tweaking website wording and advising people to rely on dedicated human-run resources rather than AI assistants. The effort underscores concerns about privacy, data collection, and the potential for delayed or denied care caused by faulty AI guidance.