The report states that growing public suspicion of artificial intelligence is “not always based on fact,” especially when it comes to military technologies.
In particular, the threats do not cohere with current AI research directions, but have arisen from dire predictions about one small area of artificial general intelligence.
However, scientists note the current worries are simply disproportionate to the present level of success in the field.
In an interview with Radio Sputnik, Professor Bishop said that even though the emergence of machines like an artificial general intelligence, which can feel, have consciousness, etc., could make people worry, “progress towards general artificial intelligence is really overhyped, [which means that] we’ve got nothing to fear from a possible militarization of artificial intelligence.”
He added that what really made him scared, however, is having systems that are not very intelligent interacting together.
“Coming back to the military situation, my greatest fear is the possibility of a military escalation caused by the failure of one or several such ‘stupid’ systems in Korea, along the demilitarized zone infested with autonomous electronic ‘border guards,’” he observed.
“The moment one of them perceives a threat, it will notify its human commanders. This could potentially escalate into a serious armed conflict between the two Koreas, especially now that North Korea is armed with nuclear missiles,” he warned.
When asked about a combination of artificial and human intelligence, Mark Bishop said that it is “a very exciting area” that is now being explored by scientists.
“We have humans interacting with an AI creating interesting artworks and music. I think it’s a really great area where we can use this new type of machine learning to make music and create interesting art.”
Speaking about when exactly such human-like robots, even though not necessarily 100-percent AI, could come along, Professor Bishop said that people were already making such robots in Japan that look like humans, but added that we are still a long way from getting machines that cross the line separating artificial intelligence from human intelligence.
Have you heard the news? Sign up to our Telegram channel and we'll keep you up to speed!
