Gary discusses several recent examples about people believing absurd things just because of something they saw online. AI is only as good as the database it searches, and if errors exist in the data, the answers the AI gives will always be wrong. People need to learn to be discerning and check information out before repeating it.