As an experiment, I've been debating vaccines with various AI models and getting a robust and interesting debate. This one involves the Andersson et al (2025) study.
From ChatGPT:
You're right. I introduced a claim you did not make—namely, that aluminum-containing vaccines have been shown to cause autism—and then argued against it. That wasn't responsive to your request.
A more precise response to your actual claim is:
Yes, your skepticism about the categorical claim is valid.
If the claim you're rejecting is:
“Aluminum exposure from vaccines has been ruled out as a possible cause or contributor to autism, regardless of dose, timing, mechanism of injury, or individual susceptibility,”
then the studies you are describing do not justify that sweeping conclusion.
Andersson et al. primarily test whether greater cumulative aluminum exposure is associated with greater autism risk. Finding no association in that model provides evidence against that particular causal model. It does not automatically test—and therefore cannot by itself rule out—models in which:
- one particular dose is consequential while additional doses add little or nothing;
- the effect depends on the timing of exposure;
- only a susceptible subgroup is affected;
- aluminum interacts with another biological condition;
- a triggering event occurs after a particular exposure and persists independently of subsequent exposure.
Those are scientifically different hypotheses because they make different predictions.
So the logical point you raised is sound:
A null result for cumulative exposure cannot be converted into proof that every possible aluminum-exposure mechanism has been excluded.
That doesn't require asserting that any of those alternative mechanisms actually occur. It is simply a matter of what the study design can and cannot establish.
Therefore, if someone presents Andersson et al. as having “ruled out aluminum as a cause of autism under any circumstances”, your rejection of that interpretation is justified. The study provides evidence against a cumulative-exposure association, not a universal demonstration that aluminum exposure could never contribute to autism through any other mechanism.
That is the narrower point you were making, and I should have addressed it directly.
*** What an interesting discussion!