My immediate thought: the algorithm is wrong! Deciding which film to watch together isn't a matter of both parties saying which films they are willing to watch

My immediate thought: the algorithm is wrong! Deciding which film to watch together isn't a matter of both parties saying which films they are willing to watch and then picking one from the intersection of those two sets. It's a negotiation, and it could be like negotiating Brexit. Take just a simple example: suppose P1 and P2 are both willing to watch films F1 and F2, but P1 would prefer to watch F1 and P2 would prefer to watch F2. But suppose P1 knows, because of an unguarded remark that P2 made, that P2 is willing to watch F1, but P1 has managed to keep quiet about their willingness to watch F2. Then P1 can "win" the negotiation by claiming that they are only willing to watch F1. So this app needs to be upgraded to act as an agent on behalf of each user. Each user tells the app which films they are willing to watch and what their preferences are, and also what information they have managed to obtain about the preferences of the adversary (significant other). The two AI-powered agents then negotiate on behalf of the users, complete with bogus deadlines, threats to walk out and so on. A premium version of the app could take account of other household duties, long-standing grievances and so on and use them as additional "bargaining chips".