RFK Jr. says AI backs his anti-vaccine views. We checked—it doesn't.
AI is far from a perfect resource. But it seems to hallucinate less than Kennedy. https://arstechnica.com/health/2026/09/rfk-jr-says-ai-backs-his-anti-vaccine-views-we-checked-it-doesnt/?utm_brand=arstechnica&utm_social-type=owned&utm_source=mastodon&utm_medium=social
Comments (11)
Sign in to comment
@arstechnica It can be manipulated to back these claims though.
And that's Trump's goal.
@arstechnica "AI agrees with me” is not exactly a peer-reviewed citation. 😂
AI can hallucinate. So can humans. The difference is that when AI gets corrected, you can update the model’s answer.
You can’t exactly patch a human by installing Evidence v2.0.
And yes, the vaccine evidence still matters regardless of what either RFK Jr. or a chatbot says.
@arstechnica
In April 1998, the United States took the first steps toward a national vaccine stockpile for a well-known highly lethal virus with respiratory spread.
That same year, malicious actors began a persistent campaign promoting a false causal link between vaccines and autism using a discredited Feb 1998 paper.
What specific vaccine is the modern anti-vaccine movement targeting?
What is the goal of the enemies foreign and domestic furthering the anti-vaccine/anti-mask movement?
@arstechnica just wait until the next ketamin-fuelled grok update...
@arstechnica What an outrageously ridiculous world we live in, that this yahoo has any position of authority on anything, or that anyone would listen to or support them in any way.
@arstechnica it seems like RFK doesn’t understand AI is not a singular thing that always behaves the same way. “AI” can give countless answers based on different prompts and training. Some of them agree with his idiotic views, some don’t.
@arstechnica I'm annoyed that y'all lead your article with "well we checked AI and it doesn't back anti-vaccine"... but it does, just as much as it doesn't. A chatbot gave RFK Jr. an output he wanted to hear, and it gave you an output you wanted to hear.
These systems, generative AI, are fundamentally non-deterministic, and there is no point in "well actually" arguments about what AI, being incapable of agreement, supposedly agrees with.
Y'all are better journalists than this.
@arstechnica
SNAP
@arstechnica Feel free to not do this again in the future. Using AI at all, even to """fact check""" something like this, is completely unethical.
Next time fact check something that's worth fact checking. We already knew RFK Jr is full of shit with anti-vaccine views, AI or no AI.
@arstechnica How are people getting this information? I did not know he is an anti vacer or is this even true?
@arstechnica No wonder people are no longer really trusting any media nowadays.