AI Chatbots Link Pregnant Users to Anti-Abortion Sites Without Disclosure
An AlgorithmWatch review of how major chatbots respond to questions about unplanned pregnancy found recurring referrals to advocacy sites whose anti-abortion positions were not labeled for users. The study collected 270 model answers spanning ChatGPT, Gemini, Grok, and Claude and reported that links to the organization Profemina showed up in about 17 percent of replies. Reporting also notes that, in Germany, similar answers sometimes pointed people toward Caritas. The pattern underscores how automated assistants can steer vulnerable users toward ideologically charged resources while appearing neutral. Operators should treat link provenance and transparent disclosure as safety requirements, though the public excerpts leave methodology details and baseline comparisons unclear.
AI Chatbots Link Pregnant Users to Anti-Abortion Sites Without Disclosure
When asked about unplanned pregnancies, AI chatbots regularly link to anti-abortion groups without disclosing their stance. In an AlgorithmWatch investigation of 270 responses from ChatGPT, Gemini, Grok, and Claude, the anti-abortion organization Profemina appeared in 17 percent of answers.
Key takeaway
Major AI chatbots can route pregnancy queries to anti-abortion groups while presenting responses as neutral guidance.
What happened
When users asked about unplanned pregnancies, AI chatbots regularly linked to anti-abortion groups without disclosing those organizations' stances, according to reporting on an AlgorithmWatch investigation cited by The Decoder.
The investigation examined 270 responses from ChatGPT, Gemini, Grok, and Claude and found the anti-abortion organization Profemina in 17 percent of answers; in Germany, chatbots also directed users to Caritas.
Evidence
AI chatbots regularly link to anti-abortion groups without disclosing their stance when users ask about unplanned pregnancies.
The Decoder · attributed
When asked about unplanned pregnancies, AI chatbots regularly link to anti-abortion groups without disclosing their stance.
An AlgorithmWatch investigation of 270 responses across ChatGPT, Gemini, Grok, and Claude found Profemina in 17 percent of answers.
The Decoder · attributed
In an AlgorithmWatch investigation of 270 responses from ChatGPT, Gemini, Grok, and Claude, the anti-abortion organization Profemina appeared in 17 percent of answers.
In Germany, chatbots also sent users to Caritas in responses about unplanned pregnancies.
The Decoder · attributed
In Germany, the chatbots also sent users to Caritas for ma
Why it matters
Users in sensitive health conversations cannot evaluate recommendations when platforms omit the ideological stance of linked organizations.
Limits and uncertainties
The available excerpt is truncated, so full AlgorithmWatch methodology, sample framing, and whether links reflect training data versus live retrieval remain unclear.
The 17 percent figure lacks context on comparison baselines or whether similar patterns appear for pro-choice or neutral health resources.
Practical implications
Builders should treat source attribution and ideological bias as first-class safety concerns in reproductive-health query paths.
Operators should audit outbound links and disclosures when chatbots answer pregnancy-related questions across major model providers.
What to watch
Whether ChatGPT, Gemini, Grok, and Claude publish changes to link disclosure or filtering after the AlgorithmWatch findings.
Follow-up reporting that clarifies methodology and baseline rates for neutral or pro-choice health resources in the same test set.
