Linguistics, History, Robot Rights, Law & Tech
Civilisation is built on morals. Without morals, humanity ceases to be anything but another animal roaming the Earth. But as with all rules, morality must be logical, it must have a kind of internal consistency. I will not argue that Kant was correct in saying that morality is a set of fixed, categoric rules that can never be deviated from. Ethics may well be based on internal, self-evident intuition or humanity’s innate drive to maximise happiness. Whatever the basis for morality, though, there must be a fixed point of reference. If it is good to help an old lady across the road today, but tomorrow it is good to push that same old lady into the path of an oncoming car, then ethics has to tell us why. Morality is not merely a way of saying “I felt like it”. Laws of ethics are either a part of nature, like the laws of physics, or a human construct, like legal rules. Neither type of law can exist without internal consistency.
Arguing Robots should not be slaves is all good and well. Maybe the case can be made based on analogies to the wrongs of the past. Perhaps the argument is based on some divine inspiration. Perhaps the focus is pure pragmatism. The problem is that we’ve been asking the wrong question. What should or should not happen is irrelevant until it is established that something could happen. There’s no point in saying that we should not eat dodos if they’re already extinct. Arguments based on the impossible are pointless. In the same way, there’s no point in saying that robots should not be slaves if robots can never, by definition, be slaves. Only by showing that a robot could be a slave, does my previous work have any meaning.
So, what is a ‘slave’? Two possible meanings exist (courtesy of a quick internet search).
(Noun) (Especially in the past) a person who is the legal property of another and is forced to obey them.
(Verb) work excessively hard.
Both definitions seem pretty uncontroversial to me. As we’re using ‘slave’ to classify an object, only the first definition makes any sense.
Now, definitions are great because they allow you to break concepts down. Break an idea down, and you can create a set of criteria. Find something in the real world that fulfils those criteria, and the concept will correctly describe that ‘thing’. Language is beautiful in its simplicity. Given our definition, what criteria then exist for being a slave? Four appear apparent to me.
- (Especially in the past)
- A Person
- Who is the legal property of another
- And is forced to obey them [the other].
By taking each criterion in turn, we should be able to see whether a robot could ever be a slave. Let’s get started.
1. appears to be more of an observation than criteria. Notice how the entire statement was in brackets, suggesting that the writer was only providing additional information. Slavery is not an institution exclusive to the past, so neither should the noun of 'slave'. Sex trafficking, child soldiers and debt bondage, all exist today, all deserve the label of slavery. Public conceptions of slavery being relegated to the past should in no way prevent the noun being used to describe real concerns in modern times. Potential robotic enslavement should be judged by no different standard than any other form of slavery.
2. is our first real criteria. Robots are not legal people. No statute or case exists, which states that mechanical beings can be people. Currently, the law has issues recognising even apes as natural people. However, the law is not the final authority on what makes a person. A person is not a thief because the law makes them such, but because of their action's nature. The law may only choose whether to punish thieves for their actions. Gaius' view equally applies to personhood. I've already explained how the Roman ius naturale applies to modern personhood, but the argument remains the same. If personhood is viewed as an innate status, robots either possess that status or not. Law cannot bestow personhood where it does not exist, nor can the law revoke personhood from a person. What the law can decide is who or what to treat like people.
That still doesn't answer the question. Do robots fulfil the criteria of personhood? Honestly, answering would involve dissecting another definition and attempting to apply that criteria to robots. No doubt the issue is the most important thing this article will turn up, however, the complexity of the subject makes it deserving of its own post. Yes, I know, hardly the most satisfying of answers. If it weren't for the fact personhood is central to society, I'm not sure I would delay answering if criteria 2 has been fulfilled. However, lamenting on what could have been is pointless. A much better use of our time is to fictitiously assume robots can be people (at least for today) and move on to our next criteria.
3. poses more difficulties. These problems are not, however, issues with saying robots could be slaves. Currently, the law is unambiguous. Nothing prevents the ownership of a machine. The modern usage of the term ‘slave’ is actually where the issues are. Any slave must be the ‘legal property’ of another. Technically, no country on the planet formally recognises that one human being can be the property of another. Mauritania was the final country to abolish legal recognition of slavery in 1981. Denying the label ‘slave’ to the victims of sex trafficking or child servitude on a legal technicality is unjustifiable. Instead, the title ‘slave’ now refers to de facto slavery. A person is a slave if they are for all practical purposes the property of another. Slavery has been outlawed, but there are individuals out there who have still choose to break that rule. Although currently, robots can be both legal and de facto slaves, noting how slavery has developed over time demonstrates solving robotic enslavement is not as simple as waving a magic wand. Even if every country on the planet was to emancipate robots tomorrow, there would still be slaves behind closed doors.
Arguably ‘another’ should be a distinct criterion on its own. There are items which amount to legal property but which are not owned by others. The rules of occupatio only apply to ownable property that doesn't already have an owner. However, I've decided to treat 'legal property of another' as a single criterion. This is mainly out of the need for simplicity, but there is also a theoretical justification. We do not say every stick is property, even when it does not have an owner. Instead, we call the sticks potential property. Only when a specific person claims the stick does it become true property under the law.
In the same way, all robots may be potential slaves, but it is only when a specific person claims a specific robot would become a slave in any meaningful sense. Both the criteria of 'legal property' and 'another' are necessary. However, they interact in such a way that they cannot be treated independently.
4. is the last element to be considered. Are robots forced to obey others? Another tricky question, however, a question that is difficult for very different reasons than the other criteria. There's not enough data to rule one way or the other. To my knowledge, there has been no case of a robot using free will to refuse a task. Defective programming or hardware may mean a machine fails in its duty, but that is not the same thing. Conscious rebellion requires an independence most robots are yet to exhibit. Moreover, even if a robot somewhere does refuse to act, its owner must then force the machine to obey. An owner who accepts the robot's decision and doesn't press the issue may not fulfil the requirements of slavery in its strictest sense. At best, we can say robots could possibly be forced to act against its own free will. Likely, that's all we need to say that robots could be slaves, but the problem demonstrates the evidential difficulties robot rights activists face. There is simply not enough data for us to make definitive factual claims.
So there we go. When asking whether robots could be slaves, we are left with a resounding maybe. Thanks to dissecting the definition, we now know what the elements of slavery are. Unfortunately, applying those same elements to the real world is slightly more complicated. Robots could most definitely be the legal property of another, being forced to obey that other. Whether intelligent machines could be people though, well the jury's still out on that one. Somebody better figure out the answer quickly— a lot depends on it.