Most people think that a 2D game is less complex to make than a 3D game with the same amount of depth or gameplay. And mathematically, they're right; there are probably more ways to fail in 3D than there are ways to succeed in 2D. But that's assuming you have an equal amount of experience in both formats, whether lots or very little or none at all. If that's not true in your case and you know you would start with a bigger lead by doing a 3D game, then you should do that instead.
The idea is to keep it simple. You'll have limited time and resources so you want to apply most of your creativity to how your game will work and how to implement the theme. You don't want to waste any effort on figuring out something you've never seen before, so it's best to simplify what you can and try to work within constraints you already know about. You want to use what knowledge you already have, and avoid needing to pick up new tools in the middle of the jam. If you have the experience building 3D games and working with 3D-focused tools, then that experience may be more of an asset to you. If you're more comfortable making and troubleshooting and polishing a 3D game quickly, then using what you know is how you can simplify the process for yourself.
Of course, making a good and successful and perfect game is not really the purpose of participating in a game jam. Challenging yourself and building your experience is a big part of game jams for many, and it's a good motivator to try something new. If you wanted to take the opportunity to figure out some 2D game design in the months leading up to the jam and then see what you could come up with, you could probably get a lot from that experience. No need to, though; just finishing in time with a working game would be an achievement as well, especially if it's the first game jam you're participating in. That will definitely be most of my focus, that's for sure.