Wouldn't it have been more useful/conclusive if the study had another control group or two with non-AI based assistance (access to math books, access to calculator, access to human experts, etc)?
FWIW, well designed study probably would still support his premise. He at least needed a group with no help, a group with no help but books and a calculator (tools), a group with human assistance, and a group with AI assistance, and a group that does an unrelated activity during the priming period.