14-April-25: MIT researchers have further developed their PAC Privacy framework to improve the protection of sensitive user data in AI models while reducing accuracy loss, offering a more efficient technique and a template for privatizing various algorithms.