The Fear of Intelligent Machines, Survey Results

Over six thousand people have voted in the Who’s Afraid of Robots? poll (which will remain open indefinitely). Many people supported this effort (see below), propelling the entry to the frontpage of Digg, as I had hoped, which increased the number of voters and brought a heap of varying opinions on the matter. The survey’s results are quite intriguing.

The poll contained 3 possible answers:

  1. Yes, I find the idea of intelligent machines frightening
  2. No, I don’t find intelligent machines frightening
  3. I’m not afraid of intelligent machines, I’m afraid of how humans will use the technology

The third option in the poll, on fearing AI usage, has come under minor scrutiny so I’ll briefly touch on my reasons for including it. Most technology has the potential of being misused or abused, either intentionally or because of ignorance. A good example is nuclear power (bomb vs. energy provider). We can assume most people would say they fear malicious use of any technology. If we would’ve had only a yes or no answer — the results would have been obfuscated as it would have been an ambiguous mix of the fear of intelligent machines and the fear of their application.

My goal was to discern between these two, which is important, because the first is a fear of new intelligences and the latter a fear of old ones (ours). Fear of malicious use applies to all technologies, so the third option serves as a filtering mechanism for this ‘general’ fear which could have nothing to do with AI whatsoever. The result is that voters weigh which feeling is overpowering: Fear of our own follies, or fear of new kinds of intelligent beings.

For example, I fear that we might use AI technologies for malicious purposes — but — I feel that the benefits of AI dwarf this fear (see Science: Powered by Intelligence). I don’t fear intelligent machines themselves at all and thus I choose the second option.


Results

Before we continue, a fair disclaimer:

This survey was carried out in an informal manner. The interpretation presented here is the speculative work of one person alone. While giving an indication of opinions within a certain group, they do not guarantee accuracy.

The flat results are as follows:

Do you, for some reason, fear the current and/or future increase of artificial intelligence?

  • 16.7% Yes, I find the idea of intelligent machines frightening (1002 votes)
  • 27.1% No, I don’t find intelligent machines frightening (1632 votes)
  • 56.3% I’m not afraid of intelligent machines, I’m afraid of how humans will use the technology (3366 votes)

The first question we need to address is: Who are these people? Are they young Icelandic academics or old American-south churchgoers? The web connects the entire world which makes it hard to say, but we do know that the majority of the first 6000 votes came from Digg users. I was unable to find surveys on the male to female ratio of Digg users, but Google reveals many claiming 80-90% of the users are male. This, combined with information provided by The Mu Life blog on user age statistics and the Modern Life on nationality, we can loosely speculate that the majority of our voters are American males aged 17-24.

Result Pie Chart for the Robot Fear Poll

Personally, I find the results a bit surprising. Digg is a technologically oriented site and the users are most likely tech savvy. That being said, 16% of voters fearing AI is quite high. I can only imagine that this percentage would rise sharply if the poll was subjected to people with less knowledge of tech and/or science. The combined fear factor of the poll — 73% — is also too much fear to be considered healthy in my book.

I think many people don’t realize AI’s full potential; imagine AI simply as the robot that brings us beer and washes our dishes (and murders us in our sleep). As I’ve mentioned before, I don’t fear AI technologies nearly as much as I fear the lack of them. We have a lot of problems, and to solve them we need more intelligence:

We’re on the gradient fast track from automation to intelligence, and with every step of the way we’re opening up new possibilites for aiding humanity with all its troubles. By building intelligent machines we’re advancing in all scientific fields, because a machine can help us process scientific data, optimize production, minimize costs and even invent new means by which we can reach our goals. No one can claim there’s enough intelligence in the world. [...] Artificial Intelligence isn’t only an incredibly powerful tool for simulating (and thereby understanding) the vast complexity that is our minds, but, once created the simulations become a tool applicable to a high number of problems from all corners of our existence. Read more…

But, of course, there are some things that we need to consider carefully in these matters, especially regarding military and security applications. Fear can be useful to make us careful, but let’s not have fear paralyze us: The quest for knowledge and scientific advances should prevail.

In closing remarks, there are a few very good points from both sides in the comments thread of the poll article that I recommend reading.

Last but not least, this entertaining and interesting survey couldn’t have happened without the support of the following great people: KuiperCliff, Chip’s Quips, One Man’s Blog, Computer Science Canada blog, Smart Machines, the 9Rules gang, fellow colleagues at CADIA — and many individuals, friends and strangers.

No related posts detected.

12 Comments, Comment or Ping

  1. Wow, this is extremely interesting and has turned into much more than I had anticipated when I first took the poll and dugg it myself. Do you have plans for more such research type polls and posts?

  2. This doesn’t surprise me at all. However, get this poll further into the minds of people and they will shift their opinion to being afraid of intelligent machines.

  3. @Tony
    Yeah, it was a great turnout. Thanks again for your support. I might do some more surveys, sure, but I have nothing specific planned yet.

    @Sir Jorge
    Hello Jorge,

    However, get this poll further into the minds of people and they will shift their opinion to being afraid of intelligent machines.

    Do you mean that if people were to see these results it would escalate fear, despite all arguments? Or that it would shift the results from fear of humans and to AI? I’m not sure I understand your point.

  4. Glad to be of assistance! Very interesting results, although not surprising really.

    You emphasise the negativity of a prevalent ‘fear factor’, which I think is probably the most relevant conclusion to be derived from this poll. It cuts both ways – 73% of people are afraid of something that hasn’t happened yet! As you say, it’s almost certain that percentage would rise in a less technologically literate sphere than Digg.

    The only way to conquer fear is to address the source of it, directly. It also means taking responsiblity for the actions of humans, individually and in groups.

    The technology does not cause fear, people do, in themselves and each other. Education is therefore a key asset in this debate, and this poll has probably helped a lot of people in that respect. Of course, we need a much wider zone of discussion than Digg, but it’s got to start somewhere…

  5. I can only imagine that this percentage would rise sharply if the poll was subjected to people with less knowledge of tech and/or science.

    I think it might even go down, actually. People with less knowledge of science are often also more sceptical about it. Especially so when it comes to artificial intelligence – many non-scientific people are essentially dualists who don’t think that any sort of true artificial intelligence will ever be possible. (They might be more open to the dangers of narrow AI, but I doubt that narrow AI will be the first thing to come into the minds of people when you word a question the way you did.)

    As I’ve mentioned before here on Inkblot Earth, I don’t fear AI technologies, I fear the lack of them — we have a lot of problems, and to solve them we need more intelligence:

    True. Sufficiently advanced AIs would probably have the capability to solve practically any problem imaginable. Then again, the same intelligence that enables them to do that would also give them the capability to cause problems far worse than anything so far.

    An AI destroying all of humanity might solve our currently existing problems, but I’m not sure if I’d like that to happen. I’d much rather take the Friendly ones – but those won’t be developed if nobody fears the possible consequences of AIs. Therefore, I can only hope for more fear. :)

  6. Very interesting discovery and a great find! Thanks for thoughtful post. I’ll be back!

  7. Fair point, Hrafn, actually, it’s all in the words. I do think, though, that whichever way you phrased the enquiry, people are inclined to seek the worst future imaginable. That’s not your ‘fault’, it’s endemic in the West. Jeez, it’s also a huge question. Perhaps we need a huge new site to discuss unnecessary pessimism?

  8. @kuiper

    You emphasise the negativity of a prevalent ‘fear factor’, which I think is probably the most relevant conclusion to be derived from this poll. It cuts both ways – 73% of people are afraid of something that hasn’t happened yet!

    Yeah. It was very interesting to see such an high ‘fear factor’ — everybody’s heard of Terminator, fewer have heard of systems like Poseidon — which have already saved several lives.

    @Kaj Sotala

    I think it might even go down, actually. People with less knowledge of science are often also more sceptical about it. Especially so when it comes to artificial intelligence – many non-scientific people are essentially dualists who don’t think that any sort of true artificial intelligence will ever be possible. (They might be more open to the dangers of narrow AI, but I doubt that narrow AI will be the first thing to come into the minds of people when you word a question the way you did.)

    That’s a good point! Leaves me wondering. Guess the only way to really find out would be to subject more people to this kind of survey.

    True. Sufficiently advanced AIs would probably have the capability to solve practically any problem imaginable. Then again, the same intelligence that enables them to do that would also give them the capability to cause problems far worse than anything so far.

    So what would prevent them from solving those problems? And what would give them an incentive to create a solution that causes greater problems?

    I see your point, in a way, similar to how nuclear power has solved some problems but created greater ones.

    But again I bring up the point that speculating that a higher intelligence would create problems without solving them, or not being able to optimize their solutions to cause few other problems — that’s beyond me. Sure, hypothetically it could happen — it’s imaginable — but there’s no way for us to predict the “problem causing” of a higher intelligence. Especially when there’s still so much we don’t know yet about how it will can built and how it’ll function.

    Keep in mind that when humans create problems whilst solving others, it’s usually because we didn’t have the mental capacity to anticipate the side effects (or that we did it on purpose, which is usually traceable to our ignorance or choices based on feelings).

    An AI destroying all of humanity might solve our currently existing problems, but I’m not sure if I’d like that to happen. I’d much rather take the Friendly ones – but those won’t be developed if nobody fears the possible consequences of AIs. Therefore, I can only hope for more fear. :)

    I’m making the answer to (argument against) this statement a new blog entry. Thanks for the inspiration :)

    By the way, I’m a bit busy these days but I’ll probably post a new reply to the other argument thread soon. Hope this one is enough food-for-thought until then!

    @Kuiper
    Hah, yeah. We are an exceptionally pessimistic group, Westerners. I think you’re right, a whole new site would probably be needed.

    It’s especially noticable in the media. We’re drawn to tales of terror and mistakes, which propels news that cover exactly that. Which in turn incites us to read more …

    Can it be mostly traced to the fact that problems are thought provoking, drive societies and give lives meaning? Just look at my comment here — I wrote a lot more in response to something I disagreed with (Kaj’s post) than your point that I completely agreed with. ‘Nuff said?

  9. Mike

    You realize that, because you do not have a random sample, your “results” are totally meaningless?

    -Mike

  10. Hrafn

    Hey Mike. Technically, there is no such thing as “totally” meaningless results. But the poll is informal and unreliable, yes, that’s why I included a disclaimer stating exactly that.

    -Hrafn

  11. Justin

    I fear stupid people with smart machines, but I also feat stupid people with stupid machines (like chainsaws). so it comes down to a fear of those stupid humans, not the machines.

  12. Very intresting poll!

    The movies like Matrix and such scare the shit out me.

Reply to “The Fear of Intelligent Machines, Survey Results”

Please read the Terms of Use before commenting!

Basic HTML allowed (a, blockquote, strong, em)


Other ..

Think Artificial is a proud member of the 9rules blog community.