Presented By O’Reilly and Intel AI
Put AI to work
Sep 4-5, 2018: Training
Sep 5-7, 2018: Tutorials & Conference
San Francisco, CA

Executive Briefing: AI safety—Problems, state of the art, and alternatives

Allison Duettmann (Foresight Institute)
4:50pm-5:30pm Thursday, September 6, 2018
Secondary topics:  Data Networks and Data Markets, Ethics, Privacy, and Security
Average rating: *****
(5.00, 1 rating)

What you'll learn

  • Understand the complex and not very transparent field of AI safety and the key challenges, including the subissues that need troubleshooting first


Allison Duettmann offers an overview of AI philosophy and explains why traditional approaches need updating, because they pave the way for a singleton AI that will hardly be benevolent. Allison then discusses potential alternative AI safety strategies and their shortcomings and shares a brief survey of interesting problems in AI safety and what we can hope for if we get it right.

While this session focuses on a positive long-term future, most ideas should be relevant for current AI, blockchain, and computer security strategies.

Topics include:

  • Why AI safety is hard: The four categories of problems in AI safety—ethical content, technical alignment, computer security, and social coordination—and the most important controversies in each
  • State of the art: The most promising strategies to tackle each problem domain, including their shortcomings
  • Alternatives: An alternative approach to artificial general intelligence that relies on a decentralized view of superintelligences, based on a paper coauthored by Mark Miller, Christine Peterson, and Allison for the 2017 UCLA Risk Colloquium
Photo of Allison Duettmann

Allison Duettmann

Foresight Institute

Allison Duettmann is an AI safety researcher at the Foresight Institute, where she conducts research and coordinates the institute’s technical programs. Her research focuses on the reduction of existential risks, especially from artificial general intelligence. At, she keeps an index of readings, podcasts, organizations, and people that inspire an optimistic long-term vision for humanity. The index is collaborative and for everyone who wants to improve the world but doesn’t know where to start. Allison speaks and moderates panels on existential risks and existential hope, AI safety, longevity, the blockchain, ethics in technology, and more. Previously, she hosted and planned TEDx talks, panels, workshops, debates, and conferences for governments, companies, think tanks, NGOs, and the public in Germany, France, Colombia, the UK, and the US. Allison holds an MS in philosophy and public policy (summa cum laude) from the London School of Economics, where she developed a moral framework for artificial general intelligence that relies on natural language processing.