AI for social good: A behavioral science perspective on opportunities and challenges for ethical AI

AI has the potential to deliver socially beneficial outcomes to individuals, communities, and society at large. As a general purpose technology, AI can be used to understand problems, seek solutions, and make decisions to reach results potentially superior to those achieved by other means. Realizing this potential, however, hinges on understanding how AI is transforming our lives so that we can steer the deployment of AI in a direction that is preferable for our society and sustainable for our future. In this talk, I will present a series of empirical findings that illustrate, across several domains, the opportunities and the challenges that confront AI for social good. In particular, I will discuss: the psychological barriers to adopting AI in healthcare, the potential risks associated with premature deployment of AI for trust in public institutions, whether generative AI may moralize violations, and the complex relationship between AI literacy and AI receptivity.