Decision Theory and the Future of Artificial Intelligence (2020 - 2024)

Abstract

Artificial intelligence (AI) is all the rage these days. It promises many new innovations which make our lives easier. At the same time, AI involves various risks and raises a number of ethical questions. It is clear that society as a whole has to take a stance on it. The future of AI is not only an engineering challenge, but also a challenge for other societal and academic sectors, including the humanities. Our project takes up this challenge and brings the corresponding efforts at Cambridge and LMU under one umbrella. To achieve this ambitious goal, we build on well-established institutional structures on both sides -- the Leverhulme Centre for the Future of Intelligence (CFI) at Cambridge and the Munich Center for Mathematical Philosophy (MCMP) at LMU Munich.

To realize our vision, this project focuses on one important aspect of current AI research: decision theory. AI innovations, including techniques from machine learning, are increasingly used to make decisions with significant social and ethical consequences, ranging from determining the news feeds on social media to making sentencing and parole recommendations in the criminal justice system. Decision theory provides and studies the standards by which such decisions are evaluated and improved. What is a rational decision? How can we train machines to make rational decisions? What is the relationship between human decision-making and machine decision-making? How can one make machine decision-making transparent (i.e. understandable to a human agent)? Which role does cognitive science play in these developments? These are some of the urgent questions our project addresses. But we do not want to just gloss on these big questions, we also want to work on concrete projects, discuss our specific research questions with each other, engage in joint research projects (resulting in joint publications), and organize joint events such as conferences, lecture series, and podium discussions.

Projektdaten