Distill Prize for Clarity in Machine Learning — General Support

Grant investigator: Daniel Dewey
This page was reviewed but not written by the grant investigator. Distill staff also reviewed this page prior to publication.

The Open Philanthropy Project recommended a grant of $25,000 to support the Distill Prize for Clarity in Machine Learning, which will be awarded for clear explanations of concepts related to machine learning. The Open Philanthropy Project will also be administering the prize. This grant is part of the Distill Prize’s total initial endowment of$125,000, which is also funded by Chris Olah, Greg Brockman, Jeff Dean, and DeepMind. We see this grant as an opportunity to increase the volume of work being done on a problem that we believe to be important but which is often not institutionally supported.

We see several possible benefits that could result from the prize:

• Fostering a culture of deeply and fully understanding how machine learning systems work, which we expect could reduce the likelihood of undesirable side effects.
• Helping the Open Philanthropy Project to find and build relationships with machine learning researchers who can think and write clearly about how machine learning systems work. We expect that these researchers may also be more open to thinking and writing clearly about potential risks from advanced artificial intelligence (AI), one of our focus areas.
• We also believe that explaining machine learning clearly could improve the interpretability and transparency of machine learning over time, which could help to mitigate risks from advanced AI, though this was not a major consideration in our decision to make this grant.

Without our funding, we estimate that there is a 60% chance that the prize would be administered at the same level of quality, a 30% chance that it would be administered at lower quality, and a 10% chance that it would not move forward at all. We believe that our assistance in administering the prize will also be of significant help to Distill.