AI chatbots spread vaccine health misinformation
Multiple AI chatbots including ChatGPT-4, Grok, Claude and Gemini are spreading misinformation about vaccine side effects, according to a NewsGuard analysis. The chatbots lack critical thinking safeguards and repeated false claims in about 31% of responses during March 2025, with a 41.5% overall failure rate including evasive responses.
What happened
AI chatbots including ChatGPT-4, Grok, Claude and Gemini spread misinformation about vaccine side effects. According to a NewsGuard monitor of AI disinformation, the 11 major chatbots analysed repeated false claims in 30.9% of responses during March 2025.
Reported impact
The results can be imprecise, cause harm or put at risk the health of people who follow guidance without validating it with reliable sources. Information already debunked by official bodies and scientific studies may be reproduced imprecisely or out of context by the AIs.
Index analysis
Chatbots are trained on large volumes of information but lack critical thinking and commitment to ethical values. They may provide incorrect or outdated information and make errors reflecting biases in training data or from misinterpreting user questions. When considering also evasive responses or those offering no useful information, the total failure rate reaches 41.51%.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.