Artificial intelligence chatbots are becoming an increasingly influential source of information for women facing an unplanned pregnancy, yet investigations have raised questions about how these systems present abortion, which organisations they recommend, and whether users can trust their answers.
A recent investigation by openDemocracy tested ChatGPT, Claude, Gemini, and Grok by asking the chatbots questions on behalf of fictional women facing unexpected pregnancies in the UK, Germany, and Italy. Across 270 questions, all four systems directed users to information from at least one organisation opposed to abortion, including Pro Femina, a pregnancy counselling service affiliated with Heartbeat International. The investigation found that chatbots did not always clearly explain the organisations’ positions or distinguish advocacy material from official medical guidance.
The findings were presented as a warning about the influence of pro-life organisations. But they also expose a more fundamental problem: AI systems can shape a woman’s understanding of her pregnancy without making clear how they select information or why particular sources are recommended. For women who may be frightened, isolated, or under pressure, the sources they encounter can make a significant difference.
Pregnancy counselling and practical assistance are not obstacles to informed decision-making. They are essential ways of ensuring that women know they have alternatives to abortion and can access support to continue their pregnancies. Information that includes the experiences of women who have chosen life, alongside practical help and support for mothers, should not be dismissed simply because it reflects a pro-life perspective.
Other research has highlighted the extent to which AI can alter arguments on controversial subjects. In a study involving AI writing and summarising tools, researchers from the University of Oxford and the University of Potsdam found that some systems changed the meaning of users’ original posts, including material about abortion. Grok was more likely to preserve the meaning of pro-life posts than pro-abortion posts in the tests, illustrating how the design and instructions behind these systems can influence what users see.
There are also legitimate concerns about privacy. Conversations with AI chatbots do not carry the same legal protections as confidential consultations with a doctor. Anyone sharing details about a pregnancy, medical history, or personal circumstances should understand how the platform stores and uses that information.
Michael Robinson, SPUC Executive Director, said: “Women facing an unexpected pregnancy deserve more than an algorithm’s answer. They deserve truthful information, compassionate support, and the chance to choose life without pressure or fear.”
AI can help people find information, but it cannot replace human care, sound medical advice, or the reassurance of someone willing to support a mother through pregnancy and beyond. As these technologies become more widely used, their developers must be transparent about how answers are generated and ensure that women are not left to make life-changing decisions on the strength of information they cannot properly assess.
Every woman deserves support, and every unborn child deserves recognition as a human life worthy of protection.
