Nature | Column: World View

Sharing

• 
• 

• 

  • 
  • 
  • 
  • 
  • 
  • 
  • 
  • 

Intelligent robots must uphold human rights

The common fear is that intelligent machines will turn against humans. But who will save the robots from each other, and from us, asks Hutan Ashrafian¹.

24 March 2015

Article tools

• PDF
• Rights & Permissions

For the best commenting experience, please login or register as a user and agree to our Community Guidelines. You will be re-directed back to this page where you will see comments updating in real-time and have the ability to recommend comments to other users.

9 comments Subscribe to comments

1. 
   Rodney Bartlett • 2015-04-26 07:51 AM

   Hutan Ashrafian writes, "Humankind is arriving at the horizon of the birth of a new intelligent race". Could this AI apply not merely to robots but to the universe, existing at every scale from the cosmic to the quantum? This means even human consciousness would, if we can see past our egos, be nothing more than AI. To explain, may I propose an alternative to the probabilistic understanding of quantum mechanics - one using hidden variables which give exact predictions, in this case by the variables being base-2 mathematics. I want to propose an alternative to the current understanding of a probabilistic universe that originated from nothingness in a Big Bang. This alternative involves binary digits, Mobius strips and figure-8 Klein bottles (in the process, a Steady State universe will be proposed). While reading this, remember that bits are not only units of information but also pulses of energy. The information in BITS or Binary digITS is the result of electrical switching, with currents normally being either "on", usually represented by the binary digit “one” - or "off“, by “zero”. A binary digit can thus be viewed as a pulse of energy. String theory says everything's composed of tiny, one-dimensional strings that vibrate as clockwise, standing, and counterclockwise currents. We can visualize tiny, one dimensional binary digits of 1 and 0 (base 2 mathematics) forming currents in a two-dimensional program called a Mobius loop – or in 2 Mobius loops, clockwise currents in one loop combining with counterclockwise currents in the other to form a standing current. (The curving of what we call space-time sounds very strange, but I think it can actually be explained by modelling space-time’s construction on the Mobius strip that can be represented by giving a strip of paper a half-twist of 180 degrees before joining its ends.) Joining two Mobius strips (or Mobius bands) forms a four-dimensional Klein bottle. And each Klein bottle can become an observable (or “sub”) universe (figure-8 Klein bottles appear to have the most suitable shape to form subuniverses). This connection of the 2 Mobius strips can be made with the infinitely-long irrational and transcendental numbers. Such an infinite connection translates^ into an infinite number of TANGIBLE figure-8 Klein bottles which are, in fact, “subuniverses”. The infinite numbers make the cosmos as a whole* physically infinite, the union of space and time makes it eternal, and it's in a static or steady state because it’s already infinite. ^ The translation could be via photons and gravitons being ultimately composed of the binary digits of 1 and 0 encoding pi, e, √2 etc.; and matter particles [and even bosons like the Higgs, W and Z particles] being given mass by photons/gravitons interacting in matter particles’ “wave packets”. * (i.e. the cosmos beyond our 13.8-billion-year-old subuniverse, which is expanding and displacing parts of the universe beyond) Informally - if an object in space consists of one piece and does not have any "holes" that pass all the way through it, it is called simply-connected. A doughnut (and the figure-8 Klein bottle it resembles) is “holey” and not simply connected (it’s multiply connected). The universe appears to be infinite, being flat on the largest scales and curved on local scales (from far away, a scene on Earth can appear flat, yet the curves of hills become apparent up close). A flat universe that is also simply connected implies an infinite universe. So it seems the infinite universe cannot be composed of subunits called figure-8 Klein bottles (flat universes that are finite in extent include the torus and Klein bottle). But gaps in, or irregularities between, subuniverses shaped like figure-8 Klein bottles are "filled in" by binary digits in the same way that computer drawings can extrapolate a small patch of blue sky to make a sky that's blue from horizon to horizon. This makes space-time relatively smooth and continuous - and gets rid of holes, making Klein subunits feasible. The Klein bottle is a closed surface with no distinction between inside and outside (there cannot be other universes, neither a space multiverse nor a time multiverse *, outside ours – there’s only one universe). * English mathematical physicist Roger Penrose’s idea of cyclic time seems to be another version of the multiverse hypothesis. Space-time is an indissoluble union, and the traditional multiverse is focused on the spatial component while the Penrose version emphasizes the temporal (time may be nothing more than the electronic display of trillions of trillions of still states each second – what is called motion of the particles in space). Erwin Schrodinger (1887-1961), the Austrian theoretical physicist who achieved fame for his contributions to quantum mechanics and received the Nobel prize in 1933, had a lifelong interest in the Vedanta philosophy of Hinduism and this influenced Schrodinger’s speculations about the possibility of individual consciousness being only a manifestation of a unitary consciousness pervading the universe.

   • Share to Twitter Share to Facebook Share link to this comment
2. 
   Blaine Bateman • 2015-04-04 07:54 PM

   @Hutan Ashrafian--Thank you for your thought provoking comments. It might be worth taking a step back from the hypothetical situation of "artificial intelligence" and consider the already real situation of machine autonomy. I think that some of the philosophical and ethical questions easily overlap from one to the other. So, we can ask, "What are the ethics of using an autonomous killing machine?". What if it makes a mistake? Is that a bigger ethical transgression than dropping "dumb" (as in, "not smart") bombs and accidentally causing innocent casualties? I am not competent go go very far in such a dialogue, but I find it worthwhile to think about. Thanks again for your thoughts.

   • Share to Twitter Share to Facebook Share link to this comment
3. 
   Peter Marchese • 2015-03-31 09:15 AM

   I object to the term "legitimate war" war can never be legitimate!

   • Share to Twitter Share to Facebook Share link to this comment
4. 
   Mary Finelli • 2015-03-26 02:53 PM

   "Animals that exhibit thinking behaviour are already afforded rights" Aside from humans, what rights are afforded to animals? Tragically and wrongly, none. It's telling that people are so interested in protecting themselves from their creations (which in all probability will not be sentient) yet there is such little regard for -and full on carnage and enslavement of- our fellow sentient species, including to the point of extinction for increasing numbers of them.

   • Share to Twitter Share to Facebook Share link to this comment
5. 
   Blaine Bateman • 2015-04-04 07:48 PM

   Hi Mary--I would paint the line at least a bit farther towards the animals, at least in a few cases. Municipalities in the US often have laws prohibiting animal cruelty. I do agree that protections for animals are pretty minimal. I think that prevailing public sentiment and taboos regulate human behavior towards animals as much as laws. Certainly there is a significant group of US Citizens who are concerned with animal rights. It is also worth noting the societal norms are very different in different countries around the world with regard to animal rights and treatment.

   • Share to Twitter Share to Facebook Share link to this comment
6. 
   John Bashinski • 2015-03-24 06:08 PM

   There is a significant, serious AGI risks community, with a reasonably sophisticated internal discourse that's been going on for many years. That discourse has gone way, way beyond the ideas in ancient science fiction, and way, way beyond the sort of soft thinking that's in this piece. The community may be imperfect, but it addresses the issues far more meaningfully than this kind of puffery. If Nature wants to engage with these issues, perhaps Nature should get commentary from the people who've spent some time thinking about them rigorously. And by this I do not mean narrow AI researchers who dismiss the issues out of hand without themselves having given them any thought. I suggest trying MIRI (intelligence.org) as a starting point. By the way, nobody, but nobody, in the relevant community takes Asimov's laws, or any other such fuzzy, unimplementable rules, seriously. They may appear in "most discussions", but they do not appear in any discussions by serious scholars of the matter, nor do they appear in any discussions that are ever going to produce anything useful. Asimov didn't take them seriously himself; they were a literary device, and many of the stories about them are basically about how unworkable they'd be... even if you could somehow build them into an AI. Furthermore, you can't worry about rule content in isolation. Building ANY set of rules into even a near-human-level AGI is an extremely difficult problem, almost certainly even if you've solved the already tremendously difficult problem of building the AGI itself. It may very well be harder than the problem of formulating the rules themselves, and it may also put serious restrictions on their content. Not to mention other issues like conflating "sentience" with "intelligence"...

   • Share to Twitter Share to Facebook Share link to this comment
7. 
   Kim Solez • 2015-04-03 07:11 PM

   John Bashinski, with respect, this approach of author bashing and name calling is not going to get you anywhere here. Nature published this article because it contains valuable new ideas, ideas that most readers of nature find refreshing and useful, with the promise of advancing our understanding of an important area of science. The fact that this important insight comes from a medical doctor, a surgeon is particularly exciting. This is a very fair environment. Let me turn around your complaint about Nature not publishing work from MIRI and challenge you at MIRI to write and submit something to Nature good enough to be published here. I think you will find great difficulty doing that. You will learn a lot from the attempt, and it will cause you to reexamine the "internal discourse" you speak about.

   • Share to Twitter Share to Facebook Share link to this comment
8. 
   John Bashinski • 2015-04-16 04:59 PM

   First, I don't work at MIRI, don't speak for MIRI, and have never worked on any project with anybody from MIRI. Ten or fifteen years ago I had some casual contact with some of MIRI's founders, and I responded to something on their blog one time. I think I might have given them 50 or 100 bucks once. That's the extent of my association with them. And if you have an issue with MIRI, there are other places to go to talk to people who actually have something useful to say. [On edit: I also don't consider myself to be one of the relevant community of scholars. It's just that as a generally informed person I'm at least aware that such a community exists.] Second, these ideas are not new. There is absolutely nothing in that piece that wasn't said better decades ago. And most of it was then thoroughly debunked. Third, this is not a scientific paper; it's an op-ed in the general interest news section, which makes it journalism, not scientific publishing. In journalism, you're expected to SEEK OUT the new and interesting material, not wait around for people to submit it to you. If Nature wants to bring people "valuable new ideas" in this section, then Nature needs to take the trouble to go out and find them. I sure HOPE this wouldn't have been publishable in the main "journal" part of Nature.

   • Share to Twitter Share to Facebook Share link to this comment
9. 
   Ezrad Lionel • 2015-03-29 05:44 AM

   Beautiful Man.

   • Share to Twitter Share to Facebook Share link to this comment