Interested in working with me?


Artificial intelligence is a fast-growing technology that could cause massive change in the world. While AI has immense positive potential, AI and machine learning researchers estimate a ~5% chance of an extremely bad outcome from high-level machine intelligence (Grace et al., 2017). My research interests are in trying to reduce the risks from powerful AI in the long-term, since even a small probability of an extremely bad outcome could result in a far-ranging impact on humanity. As an irremediable social scientist (I love studying people), I'm interested in how social science research can be used to reduce these long-term risks from AI.

I'm accepting undergraduate research assistants at the moment, who are willing to devote 9 hours / week to research questions at the intersection of social science and reducing long-term risks from artificial intelligence. One question I’m interested in right now is the relationship between near-term and long-term concerns about AI (Prunkl & Whittlestone, 2020). How will the affiliated technical areas contribute to each other? How do researchers orient towards these concerns-- are there clusters of researchers with different motivations, and what factors shape what researchers choose to work on, if so? My research questions generally aim to understand the needs and predictions of the technical AI safety and AI governance communities. I’m open to related questions of your own devising!

To make sure we both think we're working on something we consider important, and that you'll be in a good position to make progress, the position has the following requirements:

In this position, you'll be conducting extensive literature review (both formalized papers and blog posts), and synthesizing and writing up findings. Some projects may involve collecting and/or analyzing data.

As a postdoctoral researcher at Stanford, I can provide mentoring, recommendation letters, and potentially co-authorship on a conference publication. I also expect to be able to provide funding to pay participants if we do an empirical study. (If you're a student at Stanford, you can receive PSYCH 125 course credit as well.)

Please send an email to vlgates@stanford.edu if you're interested, with an unofficial transcript and CV/résumé, and a quick note about what types of questions interest you!