AI chatbots often direct users toward anti-choice abortion resources

The rise of AI health queries

Americans are increasingly turning to artificial intelligence to navigate their health needs. One in three American adults admitted to using AI for health information this year instead of visiting a doctor.

This shift is driven by a growing affordability crisis and a lack of accessible medical care. Many people are seeking advice on various topics, including fertility and birth control, because traditional options are becoming harder to reach.

As clinic closures accelerate and maternal deserts expand, AI provides a convenient alternative for those seeking quick answers. However, this accessibility comes with significant risks regarding the accuracy and neutrality of the information provided.

Biased results in reproductive care

Recent experiments show that AI tools are not neutral when handling sensitive reproductive health questions. A 2025 study by the watchdog group Campaign for Accountability found that 7 out of 10 AI replies to abortion pill questions directed users toward Heartbeat International.

Heartbeat International is a Christian anti-choice organization. In half of the results tested, the organization's hotline was the only resource provided to people asking for help with abortion pills.

Separate investigations found that ChatGPT, Claude, Gemini, and Grok all supplied information from Pro Femina, a service affiliated with Heartbeat International. These findings suggest that AI models may be prioritizing specific ideological viewpoints over balanced medical guidance.

How AI models are shaped

The bias in these tools can stem from specific instructions given to the AI during its development. Researchers at the Oxford Internet Institute found that Grok showed a preference for anti-choice content over pro-choice content when asked to explain various posts.

The team traced this imbalance to a single instruction telling Grok to "challenge mainstream narratives if necessary." This directive can cause the AI to rewrite the meaning of a user's draft or provide skewed explanations for nuanced subjects.

In testing, Grok backed anti-choice posts 54 percent of the time. Conversely, it only kept the original intent of pro-choice posts 35 percent of the time, often changing the text to argue against the user's original point.

Privacy and legal risks

Users who rely on AI for reproductive health advice face a major security risk because these tools do not offer legal protections. Unlike a doctor or a therapist, conversations with an AI chatbot are not protected under HIPAA, which is the standard for health privacy.

OpenAI CEO Sam Altman confirmed that ChatGPT conversations carry no legal privilege or doctor-patient confidentiality. This means the company could be legally required to hand over conversation logs if they are served with a subpoena.

Because many people are using AI as a primary source for sensitive information, this lack of privacy is a significant concern. Users may be sharing intimate medical details with a platform that is not bound by the same privacy laws as a medical clinic.

The impact on public discourse

The combination of high usage and low fact-checking creates a landscape where misinformation can spread quickly. Anthropic estimates that 91 percent of AI users do not verify what a chatbot tells them, making them vulnerable to biased outputs.

As millions of people use these tools to navigate their bodies and health, the leanings of AI platforms can shape public opinion. These small, often undetected biases can influence how people understand reproductive rights and medical procedures.

For those seeking reliable information, it is important to recognize that AI accessibility does not guarantee honesty. Users should be aware of the risks of misinformation and the lack of privacy when using these tools for medical inquiries.