You want to focus on the individual user, and that makes sense. But to him/her, they're using an anti-charity because they want to increase their odds of accomplishing their goal. So when thinking about whether it is "good for the world" or "bad for the world" that people do this, we have to factor in:
1: the x% probability that the user will fail, and money will be donated, and
2: the y% chance that the user would have failed without the anti-charity option, but succeeds instead because of it.
When you consider both of these aspects, it is not at all clear that the anti-charity option is bad for the world.