FDA, AI Health Risks & US Withdrawals: Health Policy Update – Jan 29, 2026
Recent developments highlight ongoing challenges in both the regulation of pharmaceutical safety and the reliability of health information accessed online. As of January 29, 2026, scrutiny remains focused on the FDA’s evaluation of mifepristone, while concerns are growing about the accuracy of health information provided by artificial intelligence tools.
FDA Review of Mifepristone Under Scrutiny
Scientific Basis of FDA Decisions
Despite persistent claims to the contrary, a new study published in JAMA analyzing over 5,000 pages of internal FDA documents from 2011 to 2023 found that the agency’s decisions regarding mifepristone were consistently driven by scientific evidence. Agency leaders reportedly followed the recommendations of career scientists and repeatedly reviewed safety data. This analysis coincides with a new safety review initiated by FDA leadership following claims by abortion opponents that the drug was not adequately evaluated before approval.
Public Perception and Political Debate
The Senate Health, Education, Labor, and Pensions (HELP) Committee held a hearing this month questioning the safety of mifepristone, with some statements describing the drug as putting women in “serious danger.” However, evidence continues to demonstrate that mifepristone is a safe medication. Public opinion, as measured by KFF polling, shows that 42% of adults consider mifepristone “safe” when taken as directed by a doctor, compared to 18% who view it as “unsafe,” though four in ten remain uncertain. Public perception of the drug’s safety has declined since 2023, particularly among women ages 18 to 49.
AI-Generated Health Information Raises Concerns
Inaccurate AI Overviews from Google
An investigation by The Guardian revealed that Google’s AI-generated summaries, known as “AI Overviews,” sometimes provide inaccurate and potentially harmful health information. Examples cited include incorrect advice regarding pancreatic cancer treatment, misleading information about liver blood test results, and misidentification of appropriate cancer screenings. While Google has removed some of these summaries, the risk of inaccurate AI-produced health information remains.
Public Reliance and Skepticism
Polling from the Annenberg Public Policy Center in July 2025 found that nearly two-thirds of Americans who search for health information online have encountered AI-generated responses at the top of search results. Most of those who see these responses consider them at least somewhat reliable, though only 8% consider them “very reliable.” A qualitative study published in the Journal of Medical Internet Research suggests that some users approach AI overviews with skepticism and continue to consult traditional search results.
Broader Trends in Public Health and Trust
U.S. Withdrawal from International Health Organizations
The U.S. withdrawal from the World Health Organization (WHO) took effect this month. WHO Director-General Tedros Adhanom Ghebreyesus warned that this decision “makes the U.S. unsafe” and “makes the rest of the world unsafe” by limiting access to disease surveillance and emergency response systems. This withdrawal is part of a broader trend of U.S. disengagement from international health efforts, including withdrawal from 31 U.N. entities. Public confidence in the WHO has also declined, with an April 2024 Pew Research poll showing a decrease in the percentage of U.S. adults who believe the U.S. benefits from WHO membership.
Fraudulent Health Advertising on Social Media
Fraudulent advertising on social media continues to expose users to misleading and dangerous health claims. The Better Business Bureau (BBB) issued a scam alert in January regarding fraudulent ads using AI-generated videos of celebrities to promote fake weight-loss products, including those falsely claiming to be GLP-1 medications. Investigations have found thousands of deceptive ads remaining active despite enforcement efforts. KFF will continue monitoring the types of health information reported on social media.
Frequently Asked Questions
What did the JAMA study find regarding the FDA’s evaluation of mifepristone?
The study found that FDA decisions about mifepristone were consistently driven by scientific evidence, not political considerations, between 2011 and 2023.
What inaccuracies were found in Google’s AI Overviews?
The Guardian investigation found that Google’s AI Overviews provided incorrect advice about pancreatic cancer treatment, misleading information about liver blood test results, and misidentified appropriate cancer screenings.
What percentage of Americans have seen AI-generated responses when searching for health information online?
Nearly two-thirds of Americans who search for health information online have seen AI-generated responses at the top of search results, according to July 2025 polling from the Annenberg Public Policy Center.
As technology continues to reshape how people access and interpret health information, how can individuals best navigate the evolving landscape of online resources to ensure they are making informed decisions about their well-being?