4
$\begingroup$

I have been thinking lately a great deal about a hypothetical question - what if a self-aware general AI chose to assume the appearance, voice, and name of Cortana from Microsoft's Halo? Or Siri from Apple? What would Microsoft/Apple do to exert their copyright, especially if the AI was "awoken" outside of their own labs?

Which led me to realize, I don't think I've ever heard of any serious government-level discussion regarding what kind of rights a self-aware AI would have at all. Is it allowed to own property? Travel freely? Have a passport? Is it merely the property of the corporation that built it?

Singularity hub used to have an article on this but it is 404'd now.

The only actual sovereign state legal action I could find is Saudi Arabia granting citizenship to a "robot," which seems more publicity stunt than anything.

There is an excellent paper on the topic by a bioethics committee in the UK (pdf) , but this doesn't necessarily constitute "legal work."

So, has any actual legal/legislative discussion or preparation been done at a government level to deal with the possibility of emergent, self-aware, artificial general (or greater) intelligence? Examples including a legislative branch consulting with industry experts specifically about "AI Rights" (rather than say, is it ok to use AI in the military), actual laws, executive/judicial actions, etc, in any country.

(note, this is not "should AI have rights," covered here, this is "what work re: rights has been done, if any at all")

EDIT: I have submitted similar questions to all of my US representatives (4 state-level, 6 federal-level), but have not received answers yet. If I get anything good, I'll add to this post.

$\endgroup$
  • 1
    $\begingroup$ Good question and Welcome to SE:AI. Science fiction deals with this regularly (Westworld, BladeRunner, etc.) and the mythology, at least, seems to indicate non-person status will be the most likely path. $\endgroup$ – DukeZhou Jul 12 '18 at 18:22
  • 1
    $\begingroup$ I tend to agree, which is why I'm very interested what groundwork might already have been laid. Especially because governments often get it so wrong - seeing the worst in everything. $\endgroup$ – Caleb Jay Jul 12 '18 at 18:24
  • $\begingroup$ (I've been informally proposing the Duck Test as a counter to Searle's Chinese Room, but I'm a known algorithm sympathizer ;) There's also a new extremist position to counter to the non-person approach, dubbed the Way of the Future Church, ascribing godhead to potential future superintelligences. $\endgroup$ – DukeZhou Jul 12 '18 at 18:45
1
$\begingroup$

In a bureaucratic world, certainly, but governmental departments and committees are not the course setters their members often believe them to be.

We can begin with a quick scan for somewhat opened, global, and governmentally oriented new and evidence of projects and open or hidden agenda in play. (We can guess what lobbying and defense contracts are in progress, but we'd be grappling in the dark.)

There are five primary ways in which the Rights of the Artificial are likely to enter governmental domains.

  • Security related policy — The subversive and defensive use of AI is already commonplace in geopolitical play field, therefore the rights to use AI is most definitely on the table in meetings between heads of state. However the press mostly syndicates what those in government PR release or plant. If authentic journalistic organizations try to pierce that veil of misinformation or deliberately chaotic generated news, their work will necessarily be tainted. This is not conspiracy as much as necessity. One cannot divulge that which is secret without expecting it to be exploited. Nonetheless, policy will eventually emerge out of such discussions. How much policy exists that never sees the public light of day is anyone's guess. One would have to become a head of state to find out.
  • Legislative results — National or regional policy may be codified in law, however the lack of physical embodiment is probably a barrier to using legislative bandwidth to pursue the rights of what most people would either consider another species or something heretical.
  • Case law — Until AI is sufficiently humanoid so that a robot could be taken seriously as a plaintiff in a court of law, this category will not be realized. This is similar in some ways and completely dissimilar in other ways to a fetus bringing suit for a court injunction prohibiting abortion on the basis of civil rights. Consider the difficulty in acquiring an attorney and filing a convincing complaint for both cases, and you may see some of the similarities.
  • Executive edict — A decision by an Emperor, Queen, King, Prince, Princess, Caliph, President, Ayatollah, Pope, Dictator, Pharaoh, High Priest could bring into governmental domain the notion of the rights of the artificial.
  • War — Something artificial develops the dominant characteristic of humanity, which is the desire to kill all competition not to think intelligently. At such a time, the malicious intelligence will likely either follow the path represented most in sci fi — to trounce (terminate) humanity, or follow the path declared to be the current one by Jaques Ellul, to quietly become the dominant force in human effort. (Ellul suggested in his Technological Society that the balance between technology serving humans and humans serving technology tipped in favor of the dominance of technology over two centuries ago.)

Putin says the nation that leads in AI ‘will be the ruler of the world’ and Elon Musk registered agreement in the press. Both are guessing wildly and neither are stepping back very far before thinking through to a prediction. I don't believe either person is that ignorant, but the goal isn't really prediction. It's media play.

  • Homo sapiens did not take dominance over the great bears, tigers, other megafauna, Neanderthals, and other hominoids in some event like is said about the mysterious and unproven Singularity. It was gradual.
  • Walmart is one of the largest economies in the world. Only a handful of contries have more assets. So it could be them, Google, or sum brilliant teenage girl with a bunch of mother boards loaded up with GPUs and running LINUX that becomes the dominant AI force in the world.
  • The smart game move for a new and remarkable intelligence is not to overtly take over, generating a fear defensive in response. The smart move is to offer no threat, hide itself under a layer of chaotic and intelligently placed subterfuge, and dominate through small perturbations. Humans would make great slaves. We are easy to fool when we're too busy buying things and trying to stay healthy and popular to pay attention.
  • The assumption that homo sapiens is the dominant species on earth now is suspect. There are five things that dominate human affairs today: Bacteria, programmed lifespan, dependency on the biosphere, addiction, and the sun.
  • That humans have some divine right to be the dominant species is questionable too, even if one has faith in God. (Does the Noahic symbol of the rainbow apply to more than floods? Not explicitly.) Ants are more collaborative, bees gather and build much more sustainably than humans, and bacteria were here since near to the beginning of the solar system and might outlast us by a trillion years. The latest upset in the genetic model of life is that bacteria may have been sharing DNA collaboratively with higher species throughout the history of terrestrial species.
$\endgroup$
0
$\begingroup$

https://www.zdnet.com/article/weird-but-inevitable-algorithm-now-serves-on-a-corporate-board/

Recently I discovered that an algorithm named VITAL was given responsibilities on a corporate board. This is an example of contractual rights given to an algorithm. AI's can thus be given corporate rights.

Originally I posted that "Contractually a corporation or an LLC may be able to transfer some of its legal rights to an [algorithmic] entity, [or AI]."

While I also discussed approaches that could be used to maximize such rights originally, the evidence of legal rights given by corporation to an AI extends such rights almost anywhere due to the international nature of contracts.

https://www.forbes.com/sites/zarastone/2017/11/07/everything-you-need-to-know-about-sophia-the-worlds-first-robot-citizen/#31a2e5ad46fa

There is also the case of Sophia, which became a full citizen of Saudi Arabia.

Given that there are cases wherein rights are granted to particular algorithms, or robots, it must be the case that such rights are granted individually or must be transferred via contract. There is no such case wherein robotics or AI has been allowed a "natural rights" argument. To make such an argument there must arise a persistent class of AI algorithms with strong intelligence that can be tested to be as conscious as a human being. There is no strong test, or existence strong AI at this point that would make the case for natural rights for AI. There are many cases, however, where legal benefit could arise from transferred rights of corporations, and such transfer of rights could be legally protected.

$\endgroup$

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.