John Licato

As we increasingly focus on making artificially intelligent systems smarter, we might lose focus of perhaps a more important goal: How can AI help us reason better? I discuss the research projects and progress of the USF Advancing Machine and Human Reasoning (AMHR) lab, which aims to answer questions such as: - How can artificial intelligence make people better reasoners? - How can we create better artificially intelligent reasoners? - How can we advance our knowledge of logic and other cognitive-level reasoning processes in order to produce better conclusions, justifications, and arguments? The AMHR lab is devoted to not only creating smarter AI, but ensuring that these advances help improve, rather than replace, human reasoners. Think of what word processing software did for humanity: it has spell-checking, it takes care of document formatting, and allows people to focus more on the creative production of high-quality, professional documents. In other words, the combination of people and word processing software together does better than either one can do alone. We believe that one day, we can do this with logical reasoning.

Although we may not like to admit it, medical doctors, judges, professors, generals, and other highly respected professionals make lots of mistakes. As an emerging body of literature shows, they are subject to many of the same cognitive biases and mistakes of reasoning as non-professionals. Fortunately, artificial intelligence can help. In this talk, I discuss the reasons for these cognitive biases and reasoning mistakes, and reveal a mode of collaboration between human experts and AI that may reduce the frequency and severity of human mistakes.