When Google launched its AI Overviews features last year, we noted plenty of examples of false, misleading, and even dangerous information that can be contained in the official-looking answers ...