The idea of an artificial intelligence (AI) uprising may sound like the plot of a science-fiction film, but the notion is a topic of a new study that finds it is possible and we would not be able to stop it.
A team of international scientists designed a theoretical containment algorithm that ensures a super-intelligent system could not harm people under any circumstance, by simulating the AI and blocking it from wrecking havoc on humanity.
However, the analysis shows current algorithms do not have the ability to halt AI, because commanding the system to not destroy the world would inadvertently halt the algorithm’s own operations.
Iyad Rahwan, Director of the Center for Humans and Machines, said: ‘If this happened, you would not know whether the containment algorithm is still analyzing the threat, or whether it has stopped to contain the harmful AI.’
‘In effect, this makes the containment algorithm unusable.’
Scroll down for video
A team of international scientists designed a theoretical containment algorithm that ensures a super-intelligent system could not harm people under any circumstance, by simulating the AI and blocking it from wrecking havoc on humanity
AI has been fascinating humans for years, as we are in awe by machines that control cars, compose symphonies or beat the world’s best chess player at their own game.
However, with great power comes great responsibility and scientists around the world are concerned about the dangers that may come with super-intelligent systems.
Using theoretical calculations, an international team of researchers, including scientists from the Center for Humans and Machines at the Max Planck Institute for Human Development, shows that it would not be possible to control a super-intelligent AI.
The team used Alan Turning’s 1936 halting problem during the analysis, which focuses on the issue of whether or not a computer program will reach a the conclusion and answer the problem – this being the halt.
However, the analysis shows current algorithms do not have the ability to halt AI, because commanding the system to not destroy the world would inadvertently halt the algorithm’s own operations
Or if the system will forever loop looking for an answer.
In their study, the team conceived a theoretical containment algorithm that ensures a super-intelligent AI cannot harm people under any circumstances, by simulating the behavior of the AI first and halting it if considered harmful.
But careful analysis shows that in our current paradigm of computing, such algorithm cannot be built.
Based on these calculations the containment problem is incomputable, no single algorithm can find a solution for determining whether an AI would produce harm to the world.
Researchers also demonstrated that we may not even know when super-intelligent machines have arrived, because deciding whether a machine exhibits intelligence superior to humans is in the same realm as the containment problem.