I think the data scraping problem is more of an opportunity cost (they think AIs should pay them more to use their content) than a concern for the traffic they account for. If traffic, and not profit, was a problem, Wikipedia would start saying they can’t support AIs either.