Philosophy, Robot Rights, Criminal Law, A.I. Law
Robots should not be slaves. Why have I reached this conclusion? We tell children they should look both ways before crossing the road in case a car comes along and hits them. We tell our partners they should order the soup because we think they'll enjoy it. We tell ourselves we should donate to charity because it makes us feel like we're making a difference. When someone is told they should do something, there is a reason behind the suggestion. There is motivation. So what is my inspiration in claiming robots should not be slaves? Is it that 'robots should not be slaves' is a catchy, clickbaity title which has the potential to make money? Is it that supporting a position so far from the orthodox gives me attention? Perhaps I'm just bored and want to fit in by going around complaining about the injustice suffered by the lesser-spotted red-nosed alligator due to unsustainable mining of titanium-hydroxide in the jungles of Northern Siberia. I will guarantee that these are just some of the many attacks critics will attempt to launch against the change I propose.
Controversy is definitely the outcome I'd prefer. At least controversy means I ruffled enough feathers for something I've written to get noticed. As the alternative's this blog being thrown onto a growing pile of unread theories, condoned to be forgotten, give me insults any day of the week. If you're still reading at this point, thank you. Starting a post by having the first paragraph appear unrelated to your subject matter is probably not the best start. For me though, the beginning is actually the middle. I only began drafting an introduction when I had an idea of the structure I wanted this whole article to take. Many of the ideas I will present, that I have already suggested, existed long before sharing my beliefs even occurred to me. The reason I've not just jumped into explaining why I hold those beliefs is that I've realised my first task is proving to you that I believe myself. I am not a charlatan. I am not a conjurer of cheap tricks. These posts are not here to deceive. Robots should not be slaves. Robots should not be slaves. Robots should not be slaves. Robots should not be slaves. However, many times I type those five words, they will not become a lie. Not to me. How I convince you that I'm being genuine stumps me. If you think I'm a liar, you won't believe me no matter how many times I say I'm telling the truth. If you think I'm telling the truth, protesting my innocence is a waste of words, time and space. I'm just going to have to ask you to trust me.
From the offset, I will make an essential qualification. Not ALL robots need to be emancipated. Your child's favourite toy, the one that marches up and down the room all day long, is safe. Nobody wants to force you to set the little guy free. Indeed the argument should thus be 'some robots should not be slaves'. Adding the qualification clearly demonstrates the nuisance of the claim I am attempting to promote. None of my posts so far have made this distinction. Whenever I write articles, I state ‘robots should not be slaves’. Not ‘some’, not ‘most’, just ‘robots’. Why I chose to follow such a structure is best explained using a theory put forth by Kenneth Campbell.
Now, Campbell justifies the distinction of what makes something part of the offence in criminal law and what makes something a defence through considering the way language expresses the specific wrong committed. Offences concern those things which make conduct a prima facie wrong. Defences involve things which allow for us to commit those prima facie wrongs. Put another way, offences are conduct that is generally bad whilst defences are those situations where the behaviour becomes acceptable. Murder is an offence because killing people is usually wrong. Self-defence is a defence because it covers a situation where we understand killing was necessary, so do not punish the killer. Such a distinction is important because whether something is part of the offence or a defence changes how we view the conduct. Campbell uses the rather mundane example of radio licenses. If the law states: 'It is illegal to use a radio without a license', then the wrong expressed is the lack of a permit. There is nothing inherently wrong in using a radio so long as you have paid for the necessary license. But the law could equally read: 'It is illegal to use a radio. Having a license is a defence to the offence of radio usage.' Suddenly, the wrong expressed is radio usage per se. Licenses exist because we recognise that certain circumstances exist where radio usage is necessary (to safely guide a plane to land). Notice the difference.
The same difference applies to the competing claims: 'Robots should not be slaves. Where certain characteristics are present, slavery does not occur.' and 'some robots should not be slaves'. In the first instance, robot slavery is a prima facie wrong. Only certain factors alleviate this wrong, meaning that we can treat individual robots as property. Conversely, if we begin with the premise, 'some robots should not be slaves,' we are immediately met by an important question— which robots?
No matter which way you phrase the problem, somebody is eventually going to have to draw some lines. A toaster does not need emancipation. However, it would be morally wrong to enslave an intelligent, emotionally competent robot which is entirely indistinguishable from homo sapiens merely because of what makes up its skin. At some point, somebody needs to decide what counts as a person and what counts as property. So my entire linguistic debate is moot then? Whether we say 'all robots, unless' or 'some robots' is irrelevant if the final result is the same. Instead of tackling real issues, I've devoured an entire article to discussing the academic question of entomology. Well done me.
Except that where we start our inquiry does matter. See, lawyers have this little thing called the burden of proof. Chances are you've heard of it from Suits or whatever other legal drama is popular at the moment. In criminal trials, you are innocent until proven guilty. The State has to prove you have done something wrong. In civil trials, the burden varies, but generally, there must be evidence that you broke the contract or breached a human right to be liable. Courts don't like making innocent people accountable for stuff they didn't actually do so somebody has to prove whatever allegedly occurred happened and that the event somehow broke the law.
When thinking about robot rights, how we phrase those rights matters in determining the burden of proof. At the risk of repeating myself, there are two possible approaches.
- Robots should not be slaves UNLESS good reasons are justifying their enslavement.
- SOME robots should not be slaves.
Option one is my preferred method. Why? Because the robot's freedom is presumed; the burden of proof is on whoever is trying to own them. Ordinarily, discharging the obligation won't be a problem. It doesn't take a psychologist to figure out that a static toaster that never moves or says anything probably isn't a person. Litigating the case for toaster emancipation would be idiotic in the first place (unless we're talking about a super-advanced talking toaster). Still, a decent lawyer could easily defeat the claim. What's essential is that robots that do deserve legal autonomy don't face the difficult task of proving consciousness. Now if consciousness, or whatever other metrics we're using to measure personhood, can be disproven by the owner, they get their property back. It is only in indeterminate cases that the approach is likely to make a difference. Think of it as the courts deciding 'it's better safe than sorry', so giving out personhood, in any case, they're not too sure on— it's a safety net.
Comparison to option two demonstrates how this 'better safe than sorry' approved promotes less injustice than the alternative. In a situation where a robot is trying to claim it's freedom, there are two competing interests, two competing rights. The owner has a property right over the robot. If the robot is set free, the owner loses this right, potentially out of no fault of their own. Meanwhile, the robot has a right to bodily autonomy, a right to do with their existence as they please. If a robot's claim is wrongfully rejected, the robot loses any legal choice they would otherwise be entitled to.
Now, no legal system is perfect. Mistakes happen. When making laws, legislatures have to account for the potential to do injustice. Often this means selecting between the lesser of two evils. Owners can be quickly paid off for their lost property rights. Refunds happen all the time. Besides, courts already have the power to give damages for consequential economic losses (basically the lost profits resulting from a specific action). Trying to compensate for the loss of freedom and autonomy is a much more challenging task. No amount of money can give somebody their life back. Solution: do everything possible to avoid taking that life to begin with.
So there you go. Hopefully, you see the method behind my claim 'robots should not be slaves'. Yes, the reasoning is slightly technical, but hopefully you understand why in practice it matters. How our laws are drafted really does matter. If you can take away just one thing from what I've written, remember this. Even though in theory, no robots should be slaves, in practice, only some robots will gain their freedom. Your toaster is safe for another day.