Part 5/14:
Professor Franconi explains that logical AI—a style of building trustworthy AI—leans heavily on formal logic systems. These systems enable machines to verify, reason about, and ensure the correctness of arguments, rules, and software updates with mathematical certainty.
Ohad delves into the profound concepts of computational logic, decidability, and undecidability. Narrowing down language expressiveness ensures that reasoning tasks within Tau remain decidable—meaning the system can always produce answers that terminate and are correct. Striking this balance is vital: more expressive languages can represent complex ideas but risk logical contradictions and infinite reasoning loops.