WGS 2018: The Future Of Superintelligence Is Far From Clear, Says Futurist

 Superintelligence is coming and it could extinguish the human race, but there is little governments can do now to control it, philosopher and futurist Nick Bostrom said on the closing day of the sixth World Government Summit (WGS 2018) in Dubai.

Bostrom is a leading thinker on “superintelligence”, which he defined as any intelligent system radically smarter than the average human brain.

In his session entitled ‘The Paradox of Choice and Morality: How Intelligent is Artificial Intelligence?’, Bostrom said that while the field is making rapid progress, its future shape is far from clear. It could evolve as a human-machine hybrid or, more probably, it could be machine intelligence.

In the near term, he said, governments can play a role in facilitating the early conflicts arising from superintelligence, such as privacy issues around medical datasets that are subjected to analysis by artificial intelligence, or the regulation of driverless cars. For the long term, he said, we have only begun to think about how to regulate a future that includes superintelligence.

He noted that foundational research is being done on the topic at universities in Silicon Valley and Montreal, and governments could help by funding more such research. However, he said: “I think it’s still too early really for governments to get involved in a big way.”

Still, Bostrom said, it is important that when the moment comes, we are ready to control superintelligence so that it serves rather than destroys humanity.

“We might only get one shot at getting it right,” he said. He said the solution might be for superintelligent machines to use their superintelligence to solve the risks that superintelligence poses.

Nick Bostrom is Founding Director of the Future of Humanity Institute, a multidisciplinary research institute working on existential risk at the University of Oxford, and Director of the Strategic Artificial Intelligence Research Centre. He is the author of more than 200 publications, including Superintelligence: Paths, Dangers, Strategies.

His talk followed a session by Daniela Rus, Director of MIT’s Computer Science and Artificial Intelligence Laboratory, on the topic ‘AI and the Future of Work’. The two sessions were held under the theme ‘Artificial Intelligence and its Impact on Governments’.