12-Jan-2021 - Max-Planck-Institut für Bildungsforschung

We Wouldn’t Be Able to Control Superintelligent Machines

Would the AI cure cancer, bring about world peace, and prevent a climate disaster? Or would it destroy humanity and take over the Earth?

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.

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.

Facts, background information, dossiers
More about Max-Planck-Institut für Bildungsforschung
More about Max-Planck-Gesellschaft
  • News

    Cells talk at each other to specialize different functions

    During development, cells must specialize their function in a well defined timeline: formation of different tissues must be coordinated from a pile of cells. The research group led by Aneta Koseska (former Max Planck Institute of Molecular Physiology (MPI), CAESAR Bonn) has now developed a ... more

    Neandertal gene variants and Covid-19

    Last year, researchers at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, and at Karolinska Institutet in Sweden showed that a major genetic risk factor for severe Covid-19 is inherited from Neandertals. Now the same researchers show, that Neandertals also contri ... more

    Genotoxic E. coli “caught in the act”

    Escherichia coli bacteria are constitutive members of the human gut microbiota. However, some strains produce a genotoxin called colibactin, which is implicated in the development of colorectal cancer. While it has been shown that colibactin leaves very specific changes in the DNA of host c ... more

  • Videos

    Epigenetics - packaging artists in the cell

    Methyl attachments to histone proteins determine the degree of packing of the DNA molecule. They thereby determine whether a gene can be read or not. In this way, environment can influence the traits of an organism over generations. more

    Biomaterials - patent solutions from nature

    Animals and plants can produce amazing materials such as spider webs, wood or bone using only a few raw materials available. How do they achieve this? And what can engineers learn from them? more

    Chaperones - folding helpers in the cell

    Nothing works without the correct form: For most proteins, there are millions of ways in which these molecules, composed of long chains of amino acids, can be folded - but only one way is the right one. Researchers in the department "Cellular Biochemistry" at the Max Planck Institute for Bi ... more

  • Research Institutes

    Max-Planck-Gesellschaft zur Förderung der Wissenschaften e.V.

    The research institutes of the Max Planck Society perform basic research in the interest of the general public in the natural sciences, life sciences, social sciences, and the humanities. In particular, the Max Planck Society takes up new and innovative research areas that German universiti ... more