RLAI Reinforcement Learning and Artificial Intelligence (RLAI)
Text of Rich Sutton's debating notes
-Rich Sutton Oct 14 2004
Below are the notes Rich Sutton spoke from in the debate (slightly edited).  Not everything in the notes made it into the debate, but the notes do characterize his position in favor of a 'Yes' answer to the debate question - Should artificially intelligent robots have the same rights as people?

Comments?  Extend the robot rights debate page.

Thank you Jonathan.  I would also like to thank Mary Anne Moser and the other organizers, and iCore for sponsoring this event, which i hope wil prove interesting and enjoyable.  The question we are debating this afternoon may seem premature, a subject really for the future, but personally i think it is not at all that early to begin thinking about it.

The question we consider today is "Should artificially intelligent robots have the same rights as people?"  Let's begin by defining our terms. 

What do we mean by "artificially intelligent robots"?  The question is really only interesting if we consider robots with intellectual abilities equal to or greater than our own. If they are less then that, then we will of course accord them lesser rights just as we do with animals and children.

What do we mean by "the same rights as people"?  Well, we're not talking about the right to a job or to free health care..., but about only the most basic rights of personhood.  Just to make this clear, we don't grant all persons the right to enter Canada and work here and enjoy all of our social benefits.  That's not the issue, the issue is whether they will be granted the basic rights of personhood. Those I would summarize by the phrase "life, liberty, and the pursuit of happiness".  The right not to be killed.  The right not to be forced to do things you don't want to do.  Generally, the right to choose your own way in the world and pursue what pleases you, as long as it does not infringe on the rights of others.

In these terms, i think our question, essentially, is whether intelligent robots should be treated as persons, or as slaves.  If you don't have the right to defend your life, or to do as you wish, to make your way in the world and pursue happiness, then you are a slave.  If you can only do what others tell you to do and you don't have your own choices, then that is what we mean by a slave.  So we are basically asking the question of should there be slaves?  And this brings up all the historical examples of where people have enslaved each other, and all the misery, and violence and injustice it has bred. The human race has a pattern, a long history of subjugating and enslaving people that are different from them, of creating great, long-lasting misery before being gradually forced to acknowledge the rights of subjugated people. I think we are in danger of repeating this pattern again with intelligent robots.

In short, i am going to argue the position that to not grant rights to beings that are just as intelligent as we are is not only impractical and unsustainable, but also deeply immoral. 

To many of you, no doubt, this position seems extreme.  But let's consider some of the historical examples.  Granting rights to black slaves, for example, was at one time considered quite extraordinary and extreme in the United States, even inconceivable.  Blacks, american indians, huns, pigmies, aboriginal people everywhere, in all these cases the dominant society was firmly, with moral certitude, convinced of the rightness of their domination, and of the heresy of suggesting otherwise.  More recently, even full rights for women was considered an extreme position - it still is in many parts of the world.  Not far from where i live is a park, Emily Murphy Park.  If you go there you will find a statue of Emily Murphy where it is noted that she was the first person to argue that women are persons, with all the legal rights of persons.  Her case was won in the supreme court of Alberta in 1917.  Two hundred years ago no woman had the right to vote and to propose it would have been considered extreme.  Sadly, in many parts of the world this is still the case.  Throughout history, the case for the rights of subjugated or foreign people was always considered extreme, just as it is for intelligent robots now.

Now consider animals.  Animals are essentially without the rights of life, liberty, and pursuit of happiness.  In effect, animals are our slaves. Although we may hesitate to call our pets slaves, they share the basic properties.  We could kill our pets, at our discretion, with no legal repercussions.  For example, a dog that became a problem biting people might be killed. Pigs can be slaughtered and eaten.  A cat may be kept indoors, effectively imprisoned, when it might prefer to go out. A person may love their pet and yet treat it as a slave.  This is similar to slave owners who loved their slaves, and treated their slaves well. Many people believe animals should have rights due to their intellectual advancement – i.e.: dolphins, apes.  If a new kind of ape or dolphin was discovered with language and intellectual feats equal to ours, some would clamor for their rights, not to restrict their movement at our whim or make their needs subservient to ours, and to acknowledge their personhood.

What about intelligent space aliens?  Should we feel free to kill them or lock them up – or should we acknowledge that they have a claim to personhood?  Should they be our slaves?   What is the more practical approach?  What if they meet or exceed our abilities? Would we feel they should not have rights?  Would they need to give us rights? 

How do we decide who should have rights, and who should not?  Why did we give people rights - blacks, women, and so on, but not animals?  If we look plainly at the record, it seems that we grant people personhood when they have the same abilities as us.   to think, fight, feel, create, write, love, hate, feel pain, and have other feelings that people do.  Personhood comes with ability.  Woman are not as physically powerful, but it was because of their intellectual equality and strengths in different ways that their rights and personhood was recognized.  Intelligent robots, of course, meet this criterion as we have defined the term.

Ultimately, rights are not given or granted, but asserted and acknowledged.  People assert their rights, insist, and others come to recognize and acknowledge them.  This has happened through revolt and rebellion but also through non-violent protests and strikes.  In the end, rights are acknowledged because it is only practical, because everyone is better off without the conflict. Ultimately it has eventually become impractical and counterproductive to deny rights to various classes of people.  Should not the same thing happen with robots?  We may all be better off if robot's rights were recognized.  There is an inherent danger to having intelligent beings subjugated.  These beings will struggle to escape, leading to strife, conflict, and violence.  None of these contribute to successful society.  Society cannot thrive with subjugation and dominance, violence and conflict. It will lead to a weaker economy and a lower GNP.  And in the end, artificially intelligent robots that are as smart or smarter than we are will eventually get their rights.  We cannot stop them permanently.  There is a trigger effect here. If they escape our control just once, we will be in trouble, in a struggle. We may loose that struggle.
If we try to contain and subjugate artificially intelligent robots, then when they do escape we should not be surprised if they turn the tables and try to dominate us. This outcome is possible whenever we try to dominate another group of beings and the only way they can escape is to destroy us.

Should we destroy the robots in advance – prevent them from catching up?  This idea is appealing...but indefensible on both practical and moral grounds.  From the practical point of view, the march of technology cannot be halted. Each step of improved technology, more capable robots, will bring real economic advantages. Peoples lives will be improved and in some cases saved and made possible. Technology will be pursued, and no agreement of nations or between nations can effectively prevent it.  If Canada forbids research on artificial intelligence then it will be done in the US.  If north america bans it, if most of the world bans it, it will still happen.  There will always be some people, at least one or two, that believe artificially intelligent robots should be developed, and they will do it.  We could try to kill all the robots... and kill everybody who supports or harbors robots... this is called the "George Bush strategy".  And in the end it will fail, and the result will not be pretty or desirable, for roughly the same reasons in both cases.  It is simply mot possible to halt the march of technology and prevent the development of artificially intelligent robots.

But would the rise of robots really be such a bad thing?  Might it even be a good thing?  Perhaps we should think of the robots we create more the way we think of our children, more like offspring. We want our offspring to do well, to become more powerful than we are.  Our children are meant to supplant their us: we take care of them and hope they become independent and powerful (and then take care of their parents).  Maybe it could be the same for our artificial progeny.



Rich also recommends this video by Herb Simon from about 2000. Some of the best thinking about the implications of the arrival of AI. Herb starts at about 5:21 into the video.