>>511345254Definitely not marginally. All tests are trainable and that's exactly how 'AI' works it takes neural networks and has them compete in a certain game until it gets better and better. The only problem with it is that it takes an immense amount of data while humans take very little.
https://pmc.ncbi.nlm.nih.gov/articles/PMC7709590/
>When standard deviations of the initial test and 2nd retest were pooled in the calculation of the effect sizes, the experimental group’s performance was 10 IQ points higher on average than that of the control group. Further, with the properly defined measures of fluid and crystallized intelligence, the experimental group showed a 15 IQ points higher increase than the control group.This is exactly how AI works it keeps running the same tests and improves its dataset. You now have AI speedrunning games, not TAS, actual AI. It works exactly the same for IQ, your one example of taking a completely untrained module that's basically a glorified search engine and expecting it to be perfect is the same thing as taking a random person and expecting them to be a competent engineer. Did they not make you read Timaeus in school? Remember when Plato takes the slave boy and makes him do geometry, the boy screws up and Plato trains him how to double a square by just asking questions? Even though he screws up constantly he eventually gets the answer if the subject is pressed enough.