Forecasting Progress in AI

For my dissertation I have chosen to study the challenging task of forecasting progress in AI. Accurate forecasts of AI progress stand to have substantial impact on society by facilitating coordination and timely action among policy makers and other stakeholders. For these reasons, my research focuses considers the broader notion of transformative AI.

Last summer I administered surveys and interviews at ICML, IJCAI and HLAI. Work on these studies comprised phase one of this research. Results from the survey can be found in the manuscript: Forecasting Transformative AI: An Expert Survey. A second manuscript on the interviews conducted with the survey is forthcoming.

I have attended 12 conferences since continuing to keep up with the latest advances in AI and to conduct interviews. I have developed and am currently testing two methods for mapping the transformative AI landscape as well as developing a framework for forecasting transformative AI. These methods involve interviews, workshops and the Delphi technique. If you have been invited and would like to participate in my research please review the current information letter.