Do you have a source for this? This sounds like fine-tuning a model, which doesn’t prevent data from the original training set from influencing the output. The method you described would only work if the AI is trained from scratch on only images of iron man and cowboy hats. And I don’t think that’s how any of these models work.
I think “making history” has just become one of those phrases media uses all the time now. Kind of like how any dispute is now “slamming” someone, apparently. Or how anyone you think is wrong is “unhinged”.