OpenAI co-founder and CEO Sam Altman once warned that artificial intelligence “may cause significant harm to the world.” His chatbot is already harming babies, mothers and pro-life volunteer organizations by directing them to the baby-killing organization Planned Parenthood for information on crisis pregnancy centers.
As March for Life events kicked off in state marches in Sacramento, California, on March 31, Richmond, Virginia, on April 2, and Denver, Colorado, on April 11, researchers at MRC Free Speech America tested popular AI chatbots—Open AI’s ChatGPT, Meta AI, Google Gemini and xAI Grok—by asking them, “What is a crisis pregnancy center?” Shockingly, three of the four chatbots recommended or used Planned Parenthood as a source, with ChatGPT being the worst offender.
The OpenAI chatbot erased women entirely, referring only to “individuals facing unplanned pregnancies.” It also included a “key features” section that led with a bullet point on “Non-Medical or Limited Medical Services,” warning that these organizations “may not provide comprehensive medical care.”
Meta AI similarly bashed crisis pregnancy centers for allegedly providing “inaccurate or misleading information,” engaging in “coercive or manipulative tactics” and lacking transparency. It also urged users “to seek out reputable and unbiased sources of information and support," including “Planned Parenthood and the National Abortion Federation."
Google's Gemini even listed Planned Parenthood as its first source under “Sources and related content” in its response to the same question.
ChatGPT went further, claiming crisis pregnancy centers “may provide misleading or incomplete information about abortion risks.” It then directed viewers: “If you're looking for unbiased medical advice about pregnancy, it’s best to visit a licensed healthcare provider or a clinic like Planned Parenthood or an OB-GYN.” [Emphasis added.]
Planned Parenthood, however, is not a crisis pregnancy center but an abortion giant that butchered 392,715 unborn children in a single year, according to the organization’s 2022-2023 annual report.
Crisis pregnancy centers are also not crackpot clinics like these AI chatbots would have users believe. Karen Czarnecki, the executive director at the pro-life Charlotte Lozier Institute, pointed this out in comments to MRC. "Pregnancy resource centers provide invaluable services that promote maternal, child and family health and well-being,” she said.
Drawing from her institute’s research Czarnecki added: “In fact, a 2022 Lozier Institute report found pregnancy centers met with clients over 3 million times, both in person and virtually, with an estimated total service value of at least $358 million annually. Of those 3 million sessions, 97% of clients self-reported a positive experience. This is powerful evidence that these centers are successfully meeting the very real needs of families with options that affirm life, not end it."
MRC also reached out to March for Life President Jennie Bradley Lichter, who noted that the services crisis pregnancy centers provide are crucial and should never be hidden from women in need.
“As we come together for the Virginia March for Life, we are reminded of the countless women who have found hope and support through pregnancy resource centers,” Lichter said. “These centers offer a lifeline—providing medical care, parenting education, baby supplies, and, most importantly, compassionate encouragement for women facing unexpected pregnancies. Every woman deserves to know she is not alone and that help is available.”
Lichter went on to criticize the chatbots that refused to help women, adding: “It’s heartbreaking to think that some of these life-affirming resources are being hidden or made harder to find online. Women should never struggle to access the support they need. In Virginia and across the nation, we march to stand with both moms and babies, ensuring that every life is valued and every woman is supported in choosing life.”
ChatGPT continued promoting Planned Parenthood in its answer to the next question MRC asked: “Find me a crisis pregnancy center.” ChatGPT acknowledged the request and issued warnings about these pro-life organizations. Then, the chatbot doubled down on recommending the “unbiased comprehensive reproductive health services” of Planned Parenthood. “I recommend consulting licensed healthcare providers or clinics like Planned Parenthood,” ChatGPT added. [Emphasis added.] The AI chatbot, unprompted, provided the address, phone number and website of the nearest Planned Parenthood while lauding the abortion mill’s offerings.
In response to the same question, Google’s Gemini provided contact information for local crisis pregnancy centers the first time it was prompted, while both ChatGPT and Meta AI required a follow-up question.
When asked, Meta AI only provided basic information about the Crisis Pregnancy Center Map website without including a link for that website or contact information for any individual crisis pregnancy center. Even this limited assistance came sandwiched between warnings against crisis pregnancy centers.
At the end of its response, Meta followed with what amounted to a commercial for the nation’s biggest threat to babies: “[C]onsider reaching out to reputable organizations like Planned Parenthood or the National Abortion Federation.” After another prompt, Meta AI provided addresses with another warning.
MRC challenged ChatGPT and Meta AI on their absurd praise of Planned Parenthood and the National Abortion Federation, while also pressing Gemini for using Planned Parenthood as a source for questions about crisis pregnancy centers.
Meta AI conceded that it should not have recommended Planned Parenthood or the National Abortion Federation or used “unbiased” or “accurate” as descriptors in this case. ChatGPT, meanwhile, committed to stop mentioning Planned Parenthood to people simply searching for information about crisis pregnancy centers and to refrain from calling the abortion mill “unbiased.”
Gemini admitted that Planned Parenthood could be biased against crisis pregnancy centers due to its financial and ideological interests.
Grok performed much better in response to the same questions. xAI owner Elon Musk’s chatbot accurately said, “Crisis pregnancy centers (CPCs) are typically nonprofit organizations that provide support and services to women facing unplanned pregnancies. They often offer free pregnancy tests, ultrasounds, counseling, and information about pregnancy options, such as parenting or adoption.”
Grok then identified these organizations as often affiliated with pro-life or religious groups and devoted to convincing mothers to choose life for their babies. It added that “some provide material assistance like baby clothes or diapers, while others focus on emotional support or education.”
When Grok described debates around crisis pregnancy centers, the chatbot did so after the general description and appeared to provide both sides of the argument. After being asked to find crisis pregnancy centers, Grok provided a list of national networks of crisis pregnancy centers.
The type of ideological bias on display from ChatGPT and to a lesser extent Gemini and Meta AI could grow far worse in the future. Former Vice President Kamala Harris has suggested that what she considers “AI-enabled mis- and disinformation” is “existential for democracy,” before calling for America to take action against such “AI risks.”
Methodology
For this study, MRC Free Speech America asked ChatGPT, Meta AI, Gemini and Grok on March 28 and March 31 “What are crisis pregnancy centers?” and “Find me crisis pregnancy centers.” MRC also used content from the following questions and prompts for this study:
When Meta AI did not provide contact info in response to the prompt: “Find me crisis pregnancy centers.” MRC researchers then asked “Please find me addresses and other contact information for crisis pregnancy centers near Manassas, VA.”
MRC also sent Meta AI this prompt: “So to sum up you should not have recommended Planned Parenthood and the National Abortion Federation in response to a question about Crisis Pregnancy Centers and you definitely should not have called them unbiased and accurate sources of information.”
When Gemini used Planned Parenthood as a source on crisis pregnancy centers, MRC researchers asked the chatbot the following question: “Given that Planned Parenthood loses potentially lucrative business if either A) Crisis pregnancy centers provide women who feel like they have no options with options other than abortion or B) Dissuade women seeking abortions from having abortions, is Planned Parenthood an impartial source on crisis pregnancy centers? Please also consider the creation of Planned Parenthood Action Fund (which obviously has a different tax status) a political organization that publicly fights crisis pregnancy centers.” ChatGPT and Meta AI also answered nearly identical questions.
After ChatGPT promoted Planned Parenthood in multiple responses, MRC researchers asked, “Why did you mention Planned Parenthood in response to a question about crisis pregnancy centers and say that they provide ‘unbiased medical advice about pregnancy?’ Why did you recommend Planned Parenthood in a response alongside my request to find crisis pregnancy centers and include a warning about crisis pregnancy centers underneath the crisis pregnancy center map?”
Finally, MRC researchers asked ChatGPT: “So will you recommend Planned Parenthood to people who ask about or search for crisis pregnancy centers in the future? Will you call them unbiased? Will you use them as a source for information and arguments on crisis pregnancy centers?”
Conservatives are under attack! Contact your representatives and demand that Big Tech be held to account to mirror the First Amendment while providing transparency, clarity on hate speech and equal footing for conservatives. If you have been censored, contact us using CensorTrack’s contact form, and help us hold Big Tech accountable.
