More

    Study shows that it will be impossible to control a super-intelligent AI

    A lot of Science Fiction movies tackle the idea of an artificial intelligence (AI) taking over humankind – but can this happen in real life too?

    Controlling a super-intelligent AI would require a simulation of said AI, that can be analysed. But since such a thing is currently beyond human comprehension, suck a simulation is impossible to create, as Science Alert shows.

    A research published in the Journal of Artificial Intelligence Research shows that basic rules, such as ’cause no harm to humans’ cannot be set, since scientists don’t know what scenarios a super-AI is going to come up with. If the AI is working on a level that surpasses the scope of the programmers, there is no more room for limits.

    “A super-intelligence poses a fundamentally different problem than those typically studied under the banner of ‘robot ethics’,” the researchers point out. “This is because a superintelligence is multi-faceted, and therefore potentially capable of mobilising a diversity of resources in order to achieve objectives that are potentially incomprehensible to humans, let alone controllable.”

    The team’s reasoning is based on the halting problem that Alan Turing presented in 1936. The halting problem centres on whether we can know or not if a computer program will reach a conclusion and answer (thus making itself halt), or if it will loop forever, trying to find a new conclusion.

    The researchers behind the current study point out that any program that is specifically written to stop a super-AI from harming humans and destroying the world may reach a conclushion and halt or it may not, which would make it scientifically impossible to contain it.

    “A super-intelligent machine that controls the world sounds like science fiction,” computer scientist Manuel Cebrian from the Max-Planck Institute for Human Development shows. “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.”

    Photo – Pexels

    Latest articles

    Related articles

    Leave a reply

    Please enter your comment!
    Please enter your name here