Asilomar Conference on Beneficial AI

• 1. Introduction

  ›

  The Asilomar Conference on Beneficial AI ran from the 5th to the 8th of January 2017, at the Asilomar Conference Grounds on the California coast. The Future of Life Institute organized the gathering, and more than 100 people showed up, not just engineers and programmers. Economists, lawyers, ethicists, and philosophers sat down together to address one shared task, formulating principles for what beneficial artificial intelligence should look like. Nobody in that room was there to build a product. They were there to write rules for an entire field before that field outran anyone's ability to steer it. What came out of those sessions, and who was willing to put their name behind it?

• 2. The 23 Principles

  ›

  A set of guidelines for AI research was the direct result, drafted as 23 separate principles rather than a single mission statement. Rather than staying inside the conference room, those principles were published as an open letter, open for anyone to read and add their name to. That format turned a private discussion among researchers into a public document other people could choose to back or ignore. How many actually chose to sign it, and who they were, tells its own story.

• 3. The Signatories

  ›

  1797 researchers working in artificial intelligence and robotics signed the open letter, and another 3923 signatures came from people outside those fields entirely. Elon Musk and Stephen Hawking put their names on it. So did Anthony D. Romero, the executive director of the ACLU, and Dutch politician Kees Verhoeven. British tech entrepreneur Tabitha Goldstaub signed too, as did American filmmaker James Barrat and Demis Hassabis, the CEO of Google Deepmind. AI researchers Ilya Sutskever, Yann LeCun, Yoshua Bengio, and Stuart Russell added their names, alongside philosophers Sam Harris and Will MacAskill. Even actor Joseph Gordon-Levitt signed, joining scientists, philosophers, and policymakers on the same page. Gordon-Levitt's name next to Yoshua Bengio's shows how far the Asilomar Principles reached beyond the labs where AI research happens.