Microsoft co-founder-turned-philanthropist Bill Gates has stated that AI, much like nuclear technology, has the potential for immense good, but only if it is used to enhance education and medicine rather than utilized in war zones.
"The world hasn't had many technologies that are both promising and dangerous," remarked Gates, speaking at Stanford's Human-Centered AI conference.
When discussing areas that AI has begun to help society so far, he stated "I won't say that there are many", but added that the technology is "a chance to supercharge the social sciences, with the biggest being education itself".
AI has the potential to overhaul the education system by identifying what works in some schools and reproducing that in others, Gates argued.
"We have not even begun to do that work of understanding motivation and engagement and teaching styles that would really improve the output of the system – better learning, less dropouts," he noted.
Gates also expressed concerns that the technology today is not as accessible as it should be.
"It is troubling that today's deep learning systems are mostly opaque," Gates said, adding that he hopes that somebody will improve their "explainability" without degrading their performance. He also voiced his belief that the technology needs to be available to more than just the large tech giants like Apple and Facebook.
Gates asserted that he was not convinced the US would dominate the AI revolution, having lost its technological edge over other countries of late.
"The US was in this totally unique position for most of these breakthrough technologies," Gates said. "Now the US is still very much the leader, but not in the same dominant, dominant way."