Will the machine become so powerful that it begins to think to a point where it is more capable than the humans? Notice, I said capable, not powerful. No doubt computers are more powerful at giving us answers faster than human brainpower, but are they more capable? so can artificial intelligence take over the world?.

  • 1
    $\begingroup$ Hi. How do you define "capable" and "powerful"? Capable of what? By the way, I think I've seen a similar question on this website before. $\endgroup$
    – nbro
    May 20 '19 at 20:19
  • $\begingroup$ This question is almost surely a duplicate, but I'm leaving it open because it generated a good answer. [See: ai.stackexchange.com/questions/tagged/ai-takeover] $\endgroup$
    – DukeZhou
    May 20 '19 at 21:27

The world seems to be operating under the "law of maximum intelligence" (I made this up), which is that the species with the most intelligence get to realise its preferences the most strongly. Put another way, the more intelligent you are, the more the world takes the shape of what you want; and this happens through your actions upon the world.

This is tautological if what you mean by "intelligence" is "ability to act well".

For example, both humans and chimps want to be happy, but you get zoos and not chimps paradise, because zoos are optimised to please humans, not chimps.

If you grant this law, then all you need is a "greater than humanity" level of intelligence, and an entity with that level of intelligence should have its preferences met.

As to whever or not this will happen, I’ll quote Sam Harris in this little talk:

All you have to grant to get your fears [about AI] up and running is that:

  • We will continue to make progress in hardware and software design (unless we destroy ourselves some other way), and
  • There’s nothing magical about the wetware we have running inside our heads, and that an intelligent machine could be build of other material. […]

Once you’ve granted those two propositions, you’ll now be hard pressed to find some hand hold with which to resist yor slide into real concern about where this is all going.

To go back to the question: yes if you have an entity/agent that is sufficiently intelligent, and with a goal that would directly or as a side-effect cause it to take over the world. Note that this is a real concern because of instrumental convergence. Good talk on this topic by Nates Soares.

I also recommend the lesswrong forums: https://wiki.lesswrong.com/wiki/AI_takeoff and Nick Bostrom: https://nickbostrom.com/superintelligence.html (or his book also titled Superintelligence)

  • $\begingroup$ I would dispute your first paragraph: it's not intelligence, it's power. There are plenty of examples in history where more advanced ("intelligent") cultures got wiped out by less advanced cultures which had more military power. $\endgroup$ May 21 '19 at 9:15
  • $\begingroup$ @OliverMason I'm talking about widely different scales of intelligences: think about the intelligence of a human vs a crow vs a fly vs a checkers program. Broadly speaking, with that kind of scale, you can consider that all humans of every culture have the same intelligence, so other factors come into play to determine which one has power. $\endgroup$ May 21 '19 at 9:19
  • $\begingroup$ Also see this post for illustration about scales of intelligence: waitbutwhy.com/2015/01/… (2nd and 3rd images about the "intelligence staircase") $\endgroup$ May 21 '19 at 9:27

Let me give you a simpler and a two letter answer No. However, not everyone may agree, as many people, including the renowned late physicist Stephen Hawking, have raised concern that “efforts to create thinking machines pose a threat to our very existence”.

In my opinion, AI is obviously a powerful technology to improve our lives. In fact, the results are there for everyone to see ranging from Apple SIRI to self-driving cars. I think all these fears are a result of Sci-Fi movies.

Everything we create is to enhance our human intelligence with artificial intelligence via machine learning and deep learning, so we can surpass the limitations of human minds.

AI is a boon to humanity and not a curse that might harm us in the future or take over the world

Source: Will AI take over humans one day? (Quora)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.