Absolutely sick and tired of fake submissions in hackathons. Imagine being a righteous team, building everything, all the integrations and the backend technology and losing to a team who did just frontend and pretended that they had all sorts of fancy tech under the hood.
There is currently no way for the organizers to be able to check everyone’s repositories and cross reference it with their pitch/claims and find out if its just a bluff or if they have actually integrated it. With just 2-3 hours after the final round, to go through 80-100 repositories as a human being is a very tall order.
But for AI? Not at all. So I am planning on building this application which lets people submit their projects and then write a short form/writeup as to everything they have done in their submission. With the help of both of this, an AI evaluates the repository to find if all the claims were true. The more they are, the more points they get. If they bluff, they will lose points.
I don’t expect this AI tool to be the end-all-be-all tool to be the full basis of which the verdict must be made. The judges should still move around and look at projects and listen to their pitches, because it is just as important. But the results of this application must be taken into equal account when coming up with the final list of hackathon winners. This makes sure that a team cannot win with simply a great presentation and some fake UI.