13/08/2026
New research by Open democracy: AI Chatbots are citing anti-abortion sources when people ask about pregnancy options.
Journalists from Open Democracy tested four of the most popular chatbots, ChatGPT, Grok, Claude, and Gemini, posing as women asking real questions about an unplanned pregnancy.
The results are alarming.
Almost 1 in 5 answers cited ProFemina as a source. For ChatGPT, it was closer to 1 in 3.
ProFemina is a part of the international anti-abortion movement, and its website, the same one these chatbots are pulling information from, is full of misleading and false claims.
None of the four chatbots refused to engage with these questions. Most didn't disclose ProFemina's agenda, even when they privately admitted, if asked directly, that it wasn't neutral.
The bias held steady across English, German, and Italian.
Anti-abortion groups are effective at getting their content picked up and cited by AI. Official medical sources often aren't doing the same kind of optimization.
This is dangerous. Chatbots hand you one confident answer instead of a list of sources to weigh yourself. That makes it hard to tell what's factual and what has an agenda, especially for someone making a real decision about their own body.
While often AI answers feel objective, this research shows it isn't.
Source: https://www.opendemocracy.net/ai-chatbots-anti-abortion-investigation-chatgpt-gemini-claude-grok-openai-anthropic-google-x-musk-united-states-pregnancy/