Endowing AI with noble goals may not preventunintended consequences.
Credit: Iyad Rahwan
New Findings From Theoretical Computer Science
We are fascinated by machines that can control cars, compose symphonies, or defeat people at chess, Go, or Jeopardy! While more progress is being made all the time in Artificial Intelligence (AI), some scientists and philosophers warn of the dangers of an uncontrollable superintelligent AI. 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 superintelligent AI. The study was published in the Journal of Artificial Intelligence Research.
Suppose someone were to program an AI system with intelligence superior to that of humans, so it could learn independently. Connected to the Internet, the AI may have access to all the data of humanity. It could replace all existing programs and take control all machines online worldwide. Would this produce a utopia or a dystopia? Would the AI cure cancer, bring about world peace, and prevent a climate disaster? Or would it destroy humanity and take over the Earth?
Computer scientists and philosophers have asked themselves whether we would even be able to control a superintelligent AI at all, to ensure it would not pose a threat to humanity. An international team of computer scientists used theoretical calculations to show that it would be fundamentally impossible to control a super-intelligent AI.
“A super-intelligent machine that controls the world sounds like science fiction. But there are already machines that perform certain important tasks independently without programmers fully understanding how they learned it. The question therefore arises whether this could at some point become uncontrollable and dangerous for humanity”, says study co-author Manuel Cebrian, Leader of the Digital Mobilization Group at the Center for Humans and Machines, Max Planck Institute for Human Development.
Scientists have explored two different ideas for how a superintelligent AI could be controlled. On one hand, the capabilities of superintelligent AI could be specifically limited, for example, by walling it off from the Internet and all other technical devices so it could have no contact with the outside world – yet this would render the superintelligent AI significantly less powerful, less able to answer humanities quests. Lacking that option, the AI could be motivated from the outset to pursue only goals that are in the best interests of humanity, for example by programming ethical principles into it. However, the researchers also show that these and other contemporary and historical ideas for controlling super-intelligent AI have their limits.
In their study, the team conceived a theoretical containment algorithm that ensures a superintelligent 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.
“If you break the problem down to basic rules from theoretical computer science, it turns out that an algorithm that would command an AI not to destroy the world could inadvertently halt its own operations. 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”, says Iyad Rahwan, Director of the Center for Humans and Machines.
Based on these calculations the containment problem is incomputable, i.e. no single algorithm can find a solution for determining whether an AI would produce harm to the world. Furthermore, the researchers demonstrate that we may not even know when superintelligent machines have arrived, because deciding whether a machine exhibits intelligence superior to humans is in the same realm as the containment problem.
The study “Superintelligence cannot be contained: Lessons from Computability Theory“ was published in the Journal of Artificial Intelligence Research. Other researchers on the study include Andres Abeliuk from the University of Southern California, Manuel Alfonseca from the Autonomous University of Madrid, Antonio Fernandez Anta from the IMDEA Networks Institute and Lorenzo Coviello.
The Latest Updates from Bing News & Google News
Go deeper with Bing News on:
- Premature Freakouts about Techno-Enhancementon January 23, 2021 at 9:09 am
How can we protect privacy in an age of tech-enabled telepathy?” We expect boosters of techno-enhancement to minimize technical obstacles. Transhumanists can’t wait to become super-intelligent, ...
- New Findings from Max Planck Institute say that it is Impossible to control Superintelligent AIon January 21, 2021 at 3:15 am
To counter Threats of superintelligent AI to humanity, Nick Bostrom had proposed using containment algorithm. But recent report from Max Planck Institute reveal that it may be impossible to build one ...
- Artificial Intelligence: Cure for What Ails Us, or Looming Threat to the World?on January 19, 2021 at 7:14 am
Led by Manuel Alfonseca, the authors claim that based on computability theory, a superintelligent AI cannot be contained, and thus poses a threat to all of us. Kriegman’s team calls their biological ...
- Computer Scientists Wouldn’t Be Able to Control Super-Intelligent Machineson January 14, 2021 at 9:02 am
Computer Scientists Wouldn’t Be Able to Control Super-Intelligent Machines. We are fascinated by machines that can drive cars, compose ...
- AI AI Captain! New Study Shows Humans Can’t Control Superintelligent Machineson January 14, 2021 at 1:32 am
AI is smarter than humans at specific tasks such as playing chess or solving the Schrödinger’s equation. At this rate, AI might end up as a force beyond human control.
Go deeper with Google Headlines on:
Go deeper with Bing News on:
- Hitting the Books: AI doctors and the dangers tiered medical careon January 23, 2021 at 8:33 am
Frank Pasquale’s 'New Laws of Robotics' shows how the promise of faster, more efficient medical diagnoses from AIs can also be a double-edged sword, cutting off access to the quality care provided by ...
- Why seeing robots in pop culture is importanton January 22, 2021 at 6:04 pm
Chances are, the first robot you ever encountered was one from popular culture. Ever since the Czech writer Karel Čapek coined the term “robot” in his 1920 science fiction play “Rossum's Universal ...
- Jumpstart your career in data science with this six-course training bundleon January 19, 2021 at 3:02 am
Business Intelligence & Data Science Super Bundle contains over 60 hours of content on how to drive profitable business actions using Power ...
- Researchers Calculate That Super-Intelligent AI Will Be Impossible To Containon January 15, 2021 at 6:52 am
Pretty cool, but take heed—one day it be us doing the bidding of super smart artificial intelligence (AI ... to contain such an "superintelligence" machine. "We argue that total containment ...
- It would be impossible to pull the plug on a super-intelligent machine that wanted to control the world and harm humans, scientists warn in paper on the development of AIon January 14, 2021 at 11:41 am
Physicists show it would be impossible to control AI that is out to destroy the world, because current containment algorithms would halt their own operations when commanding it to stop.