Since I anticipate somebody appealing to the Flynn effect [0] or similar improvements in intelligence-testing scores across the recent centuries, I would head them off by mentioning Spearman's hypothesis [1]. The obvious reading of the hypothesis is horribly racist, but we can take its contrapositive: If there are largely no changes between humans in recent millennia, then the differences in testing scores are due to cultural biases in the construction of the test and the education of people. And it happens to be the case, due to pedigree collapse [2], that we can be mathematically confident in the homogeneity of humanity (to say nothing of confidence from molecular biology [3]).
[0] https://en.wikipedia.org/wiki/Flynn_effect
[1] https://en.wikipedia.org/wiki/Spearman%27s_hypothesis