Take your "too broad" problems and break them down suck that at least a part might be doable within the allotted time frame.

People, especially young women, spend tons of time doing laundry around the world. Doing laundry in a time consuming and manual way keeps them from attending school, which makes them unable to rise out of poverty. This is a big issue.

Will you solve the issue of how time consuming it is for people to wash clothes by hand? No. But could you design, in a computer, a pedal-driven washing machine possible to make out of plastic? Yes. Build a prototype? Perhaps. Make instructions available? Maybe. Roughly estimate costs? Sure.

If done well, this could help fix a problem the world is facing. Or at least help create good ideas for others to work more on.

Don't limit yourselves.