How secure are computers and how does artificial intelligence impact security? In this program, Christopher Kruegel, professor of computer science at UC Santa Barbara, explores two key questions related to security and artificial intelligence. First, how AI can help to improve security. For decades, security solutions have leveraged traditional machine learning models. Not surprisingly, recent advances in AI have opened up exciting new opportunities. Second, the security of AI systems themselves. Like any other software application, they can be exploited. Given their often-critical role, it is imperative to secure AI against attacks such as training data poisoning and adversarial inputs. Series: "GRIT Talks" [Science] [Show ID: 40086]