Eliezer Yudkowsky at the 2006 Singularity Summit
Eliezer Yudkowsky is co-founder of the Singularity Institute and a key researcher in the field of AI ethics.
At the age of sixteen, Yudkowsky first read about 'intelligence explosion' - a future scenario in which an artificial intelligence improves its own intelligence until it is vastly smarter than humans. He described this discovery as causing "a vast, calm feeling that now I know how I'll be spending the rest of my life". He left school and taught himself computer science, mathematics, programming, physics, and cognitive science. In 2000, five years after that first revelation he co-founded the Singularity Institute to promote "points of leverage for increasing the probability that the advent of AI turns out positive".
Yudkowksy envisions the near-future creation of self-modifying artificial intelligences with the capacity to recursively self-improve at an explosive rate, transforming from human-level intelligence to superintelligence too quickly for humans to predict or react. To deal with this challenge, he has pioneered the field of "Friendly AI", which seeks stable goal systems that can guide such intelligences and keep them benevolently inclined toward humans.
He has since written several foundational articles on the subject, including 'Creating Friendly AI' (2001), 'Coherent Extrapolated Volition' (2004), 'AI as a Positive and Negative Factor in Global Risk' (2008), and 'Timeless Decision Theory' (2010).
Through his work with the Institute, he has co-founded the Singularity Summit, an annual TED-style event featuring prominent scientists, philosophers, and visionaries who discuss Singularity-related issues.
Yudkowsky maintains a keen interest in raising awareness of Singularity-related topics, and often writes popular articles on probability theory, decision theory, cognitive science, and artificial intelligence at LessWrong.com.