Paul Christiano, a former researcher at OpenAI, has warned that artificial intelligence (AI) poses a major risk to humanity and estimates that a 10-20 percent chance exists of an AI takeover resulting in many human fatalities. OpenAI is an artificial intelligence research and development laboratory where Christiano was a researcher on the language model alignment team.
During an appearance on the Bankless podcast, Christiano highlighted the worries that the human race may face if AI technology is developed to reach ‘human level’ thinking. In fact, Christiano suggested that the odds of a negative outcome could rise to a 50/50 chance if AI reaches human-level capability. Christiano expressed reservations about the rate of advancement in AI technology; he argued that rather than fast transformation as proposed by AI skeptics, the shift would occur slowly, followed by a more rapid acceleration over time.
These concerns echo those of many other AI professionals who recently signed an open letter begging for a six-month pause in the development of advanced AI, due to potential risks.
OpenAI is a research laboratory, founded by Elon Musk, Sam Altman, Greg Brockman, and other technology companies in 2015, with the aim of advancing digital intelligence in the way that is most likely to benefit humanity as a whole. OpenAI is dedicated to producing a safe and secure AI that is open to all. Their core mission is to ensure that AI serves humanity by benefiting us while minimizing risks. OpenAI is actively researching, building, and deploying artificial general intelligence in order to understand and develop intelligences that will help people around the world.
In contrast to OpenAI, Paul Christiano is a researcher and a PhD student at the University of California, Berkeley, with research focus in artificial general intelligence. His expertise is in technical problems of value learning, alignment, and impact measurement, specifically applied to artificial intelligence. Christiano has co-authored several papers on the potential risks posed by AI and the importance of aligning machine learning algorithms with ethical values. He is a prominent figure in the artificial intelligence ethics and safety community and currently heads the non-profit Alignment Research Center which is dedicated to mitigating the risks of unaligned AI systems.
