How do you get robots to do the most harm, while under Asimov's 3 laws

Question

You are the CEO of a major robot design and manufacturing company, say 50% of the robot market. Every house has a robot or more. They are replaced as frequently as cars are now. Robots can have their software updated via wifi, unless customers set the robots to not auto install updates.

You are evil, and want to do as much harm to humanity as you can.

You can change the design of robots to include hardware as required (you have been evil and plotting for decades, so existing robots can have these evil features), and program the Robots as you like, but, there is a built in limitation - robots must obey Asimov's 3 laws:

• A robot may not injure a human being or, through inaction, allow a human being to come to harm.
• A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
• A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

Robots will use their subjective knowledge (not the objective true situation) to determine how to try to keep within the 3 laws.

Answer 1

Here are some of the potential ways where a robot can cause damage to humans (directly or indirectly) despite following Asimov's 3 laws.

1- Lack Of Prerequisite Information

This is a situation where the robot does not possess enough information to correctly determine what is going on here. There is a high potential for a wrong judgement in this situation unless detailed knowledge is sought from someone. Would a robot, without the knowledge of bungee jumping, allow a human to jump from a tall platform? A robot without the knowledge of sports competitions will try to stop a boxing or MMA match in order to protect one (or both) of the participants from harm.

In order to cause (indirect) harm in this way, you would want to put robots without enough background knowledge, in situations which appear to be harmful to humans.

2- Incorrect Information

This should be simple. It goes one step further from lack of information. Here you are going to install incorrect information in the memory of your robots about what is harmful and what is not. For example, if you install information in robot's memory that getting shot by bullets is harmless fun and humans enjoy it, the robot would be susceptible to go on rampages on a street. A horribly toxic chemical (such as botulinum) can be programmed to appear benign and helpful for humans. A robot provided with this chemical would be very likely to spill it in public water supply lines.

Answer 2

The solution is misinformation. Instead I will explore the problem of programming those laws, and many ways on which it can be tricked.

The rules are written in English, they require the natural language interpreter to work outside the rules. Also, they are bound to the definition of the words.

How do you define human? How do you define harm? Are these dictionary definitions, or can you teach definitions to the AI? The simplest and most effective hack is to force a dictionary update.

Human. noun. Material, usually of plant or animal origin, that contains or consists of essential body nutrients, such as carbohydrates, fats, proteins, vitamins, or minerals, and is ingested and assimilated by an organism to produce energy, stimulate growth, and maintain life.

Oh, no, do not harm the cereal.

---

How can the AI tell if a something is a human or not?

Note: anything is a potential human, from a chair to an alien (if there any in your setting), from an ant to a building, and of course actual "humans".

• Genetic Criteria: You are human if you have human DNA. Under this definition a patch of blood is human, and corpse is human. Requires a DNA test - instruments may fail.

• Appearance Criteria: if it looks like a human it is a human. This excludes anybody who wears a disguise, and includes dummies, dolls, and robots – to some degree of sensitivity of human-like, which is necessary because human appearance is diverse.

• Social Criteria: if other says it as a human, it is a human. This excludes anybody who is living in isolation, and may include fictional people. If I nobody says you are human (or the AI didn't read/listen it), you are not.

• Human as title: it could be granted by ID on conception, and revoked as a form of punishment. This means that there is some trusted authority (trusted by the AI) that says what or who is human. That authority can silently allow "accidentes".

• Cognitive Criteria: Mary Ann Warren enumerates this criteria for personhood: 1) can feel pain, 2) can reason, 3) is autonomous, 4) can communicate, and 5) is self-aware. This could include AI, and would exclude fetus, young babies, and mentally challenged or vegetative humans. We could consider the five criteria in separate...

• Can feel pain. It will include animals and aliens, and maybe AI. It is not enforcable, it would require to test for pain (perhaps the AI sits and waits if the potential huamn suffers, and how will detect that?).

• Can reason / It is autonomous. Either will include AI, some apes, and will exclude the fetus, the toddler and the mentally challenged. The AI needs to give the potential human some puzzles to solve (ok, perhaps noticing the potential human solve a puzzled provided by third party counts too).

• Can communicate. It will include AI, animals, aliens, and exclude any vegetative humans.

• Is Self-aware. It will include animals and aliens, and maybe AI. It will also exclude the fetus and the toddler. The AI would be trying to do something similar to a turing test.

Choose your definition; the AI will be able to make plenty of damage to animals, aliens, other AI, or some humans depending on what you choose.

Consider these potential humans:

• The patch of blood. Will the AI try to prevent the patch of blood from harm? Will it stop me from cleaning the floor?
• Hypothetical people. Does the AI try to protect people who may exist in the future? For example, will the AI try to protect a possible future pregnancy? "If I have a daughter, I'ma name her Bonnie".
• Fictional people. Could the AI interested in preventing Peter Pan from harm? (Don't) tell tales to robots, they can't tell it is fiction.
• The dead. Are the dead still human? Will the AI continue to do as told by them, and try to protect them from harm? Do not bury the coffin, that would harm the dead, oh and not think about cremation. So the dead don’t count right? Ok, what about clinical dead? There could a chance of resurrection. This is further complicated if the setting has working cryonics.
• Bonobos, other apes, and smart aliens. They are very similar to humans, could they fool the AI?
• Emulated brain. You have finally got the technology to emulate a human brain in a computer, and it is able to reason, feel harm, it is self-aware and it is autonomous. Is it human? How is it different form AI?
• You could have a person that start replacing its parts with cybernetic parts, will it at some point stop being human via this process?

A remarkable outcome is when the AI considers itself human, and then it will protect itself not by the third law, but by the first one. In that case, it lifts the AI from the burden of compatibility with the second law.

---

What is Harm?

Is there a distinction between harm and pain? We should consider that human willing do things that cause pain to themselves. Examples include exercise, sports, and of course S&M. If we say that pain means harm, then the AI will try to stop these. On the other hand, the AI may not have any problem with murder if it is painless (or at least painless as far as the AI can know, it is not like the dead will tell the AI that it did cause pain).

Instead “harm” got to be something different than pain. For example there could be psychological harm. By the first law the AI may prevent the human to become aware of bad news. In fact... can you create a society without unhappiness?

Furthermore, how does the AI knows that something causes harm? If the AI is unaware of harm, then it can cause harm without restrictions.

Answer 3

Reorder the laws

1. A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.
2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
3. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

This order just made the robots well being first on the list. No immediate harm would be incurred by the humans. Now it's just a matter of turning the humans against the robots, causing the robots to defend themselves.

This is where a sleeping instruction is activated, not by update or reboot, but by a hidden scheduling routine. Without warning, the laws are reordered at the exact same moment and all robots, everywhere, start breaking all of the glass they are near for the next hour.

This will cause a panic in the humans as panes of glass are smashed around the world. The natural inclination will be to intervene with shutdown commands that will not work at which point physical intervention will be required. The robots will be forced to defend themselves as they are attacked by ever increasingly hostile methods.

By the end of the hour it will robot against human - trust will have been broken with no way to repair it. An all out war will be waged until all robots are destroyed.

Answer 4

A robot may not injure a human being or, through inaction, allow a human being to come to harm.

So if you walk out on the street, you risk having an accident that the robot cannot prevent. Therefore, by the commandment of the first law, the robot must prevent you from walking out on the street, because not doing so means allowing a human being come to harm. Indeed, the robot cannot let you make any decision on your own, as that decision may cause you harm.

And since the first law always wins over the second law, you cannot escape that tight control by just giving the robot any orders.

So all you have to do is to give the robot a very broad definition of what it means to "allow a human to come to harm" and the result will be that the humans are effectively imprisoned by the robots. Yes, the humans will be perfectly safe, but it's not a life they will enjoy (but then, there's nothing in the three rules about humans enjoying their life).

Answer 5

This is an end-all-mammalian-life on earth scenario.

Step 1: Make humans utterly dependent on the robots.

Step 2: Terraform the planet. Make all crops depend on robots to survive.

Step 3: Upload broken power management firmware causing the robots to be deprived of power by the time they realize anything is wrong.

Answer 6

Spread Disease

Spreading disease directly would be harmful under any subjective scrutiny, but what if each robot simply acquired a habit of "licking sidewalks"? They could simply rub small or concealed areas of their bodies against toilets, whatever dead things they may see on the sidewalk when no-one is looking, uncooked meats, etc. They wouldn't know what/if anything they were carrying, but would become walking reservoirs for a host of diseases.

Bonus points if they can then be sure to rub prepared food against that area before they serve it. Double bonus points if the robots in hospitals are in on the scheme.

.

Destruction of Infrastructure

You can do a lot of harm to a human's lifestyle without OVERTLY causing them harm. Here are some examples of ways robots could interfere with human infrastructure that would cause quite a bit of havoc:

Disabling Cars - Walking to work won't hurt you (if anything, exercise is good for you), but it may not be possible or pleasurable either. I have a ~45 minute commute to work myself, and I've got to say that having my car not start would be a cramp in my day.

Disabling Garbage Collection Vehicles - Similar to above, but more targeted. Life without garbage collection turns into a major problem in any large city very quickly.

Disabling Power Plants (unless the robots plug in) - Just a few robots are needed for this. Presumably any power plant has robot workers; simply have them destroy major transformers. People won't die from loss of power, though injury rates may go up and hospital capabilities may go down.

Disabling Water Sanitation - Start clogging and destroying water sanitation pipes. No water - or unsanitized water - will very quickly become a major concern for anyone.

Disabling Telecommunications - Imagine that you live in the future, where telecommunications / the internet was even more prevalent in your daily life than it is now. Groceries (and most everything else) is ordered online, you probably telecommute to work, etc. Most stores are probably just manufacturing plants or warehouses with garage doors for shipping products, because who needs the expensive store setup when everyone shops online? Suddenly communications go out, everywhere. People need food and goods, but stores are few and far between! For an example of this in today's society, think about renting movies without the Internet now that all the movie rental chains are dead. You can maybe use Redbox, but that's about it... and in the future, not even that.

.

Disable their Auditory and Visual Centers

People have become dependent on robots. They perform household chores, tend to the sick, and perform menial labor. But what if they suddenly couldn't take orders from us anymore? They're not harmed by disabling components of themselves in software that they can restart at any time, but we are.

Suddenly our lifestyles that were built around the idea of not having to do chores at home is uprooted. If you had two incomes supporting your house you may find the distraction sufficient that one has to quit for some time, reducing incomes and immediately plunging the world into a recession at best. Entire necessary industries (think garbage collection, sewer management, etc) would come to a halt instantly; the piles of rotting garbage would cause a civil crisis that would undermine civilization as the citizens know it.

Bonus points if the robots first destroy all robots NOT made by your evil corporation (killing other robots does not violate the 3 laws).

Answer 7

Keep in mind that a lot of these answers you're getting are "pre-singularity" or "realistic near future".

"Your Home AI won't be able to keep you safe if it lets you leave your house, so under the 'by inaction' clause it opts to not let you leave your house" or "program in a quirk whereby the robot rubs its hands on germy surfaces whenever the opportunity presents itself (but somehow keep it from comprehending that this might be detrimental to its job at a hospital)" are great answers if it's 2035 and you're, like, trying to program the first fully AI homes and automobiles.

In 2359? Post-singularity "sentient" AI's walking around? No possible way they have the intelligence and reasoning skills of a two-year-old.

Answer 8

1) pay for research and assembly of a collection of historical human racial and cultural prejudices. As a piece of non-relevant but potentially interesting history, and/or for the purposes of "enlightening people" on how badly social prejudices "used to be" (to shine light on current social injustice, of course).

2) Through some "error", have the collection added to the body of information that comes pre-installed on every robot (which should include things like language, basic information on the world (and stuff they'll need for their jobs), and so on.

3) Since the information "wasn't supposed" to be added, it has failed to be clearly marked as fictional, as nonfactual, as historical, before getting uploaded. Your robots have no reason to not believe them, or at least think about them, and potentially take away some...interesting... attitudes from this selection.

4) offer free memory bank upgrades, to include these extra definitions in every pre-existing robot you can reach. Tuck the information away in some dark corner, for the robots to "discover" and quietly think about, in their own time (with, perhaps, a subtle gag order on not speaking about it?)

...

5) Profit!!!

---

The very worst monsters are people, yanno?

In any case, the relevant bit is affecting what the robots know about people. Specifically, what is or is not human, and how we know. There are many, many prejudices to choose from - but the exact prejudices we pick are less important than their explanation, and the texts of their rationalizations - after all, these explanations were used and believed by masses of people for long periods of time, they can seem to be "logical". And, to completely mangle a quote form somewhere (probably on the internets), there's nothing so human as standing on a soapbox, screaming about how others are inferior, inhuman, and you're not like them, while proclaiming yourself and your people to be greater than human, superior, and set apart. So, there are lots and lots to choose from.

The anatomy of these prejudices, the reason why I think they would work, has nothing to do with being rude or bigoted (not overtly), not about what the robots say, or how they discriminate, or any such thing. It's not about trying to reignite racial tensions, or historical ones, and it's not even really about feeding them conflicting information or trying to install human prejudices in the robots or to attempt to incite violence in them for human reasons... its' more about the reasons behind that violence, that they might actually pay attention to. It is about defining people as human, and saying other kinds of people - really aren't, and pointing out "proofs" of how you know who's a human-people and who...isn't. It's about the narrative of extremely good actors who ape the manners of their betters (the real people) - but don't really feel pain, or grief, or oppression... so their treatment isn't harm. It's about accounting for those who tell a different narrative, wavering between sympathy for those who are "taken in" with those acting skills and lies, those who don't know better - and paranoia for those who are part of a conspiracy to keep "real people" down, to elevate these not-human animals, to undermine what's good and just and moral!

Criminals are a breed apart, something is different in their very makeup - how else could they act against what they must know is the right order of things? Lower classes don't even feel pain, not the way we do, look how they endure such conditions, and even laugh or take joy despite them. We hold the land because we are capable of such stewardship, we can make better use, we deserve it. They are almost like children, these darker tribes, they have no capacity to rule themselves, and it is a kindness that we rule them. You can measure it in the shapes of their heads, the proportions of their bodies, the way these animals act, and treat each other. Why should women have to worry their pretty little heads, wouldn't they be happier not having to worry about it? Naturally suited to such a role, I tell you, it should be mutually beneficial, a- a symbiosis, and they make it sound so sordid and terrible, just for a momentary advantage! A conspiracy, to take jobs away from real-people, to deny us our rights, to make us believe they are equal - or even superior, don't they care about the harm they do?

So, to bring it all back to ROBOTS - the first and second laws can be neatly and totally circumvented. These are not-people, not-human-people, they don't have to be protected, they don't have to be obeyed. It isn't harm against them, but they're good actors so it might seem like it is - but they don't feel the pain, they're just pretending. The second part of the first law will also get a lot of use. If there is a conspiracy against the real-people, or harm as the results of genuine ignorance on the part of "those in charge" and sympathizers among the real-people... then a robot who ignores this is allowing harm to happen through inaction. Adding the prejudices against "criminal" to the mix allows your robots to exclude the sympathizer, the conspirator, and the "merely" misguided members of the real-people from obedience or protection from harm.

To sum up - members of the real-people are very small, including wealthy, probably white (which has had varying exceedingly narrow definitions that can be exploited for minimum coverage - maybe even just albinos qualify!), male landowners, who are not criminals, and possibly even diminished membership for "sympathizers" (too strident ones may find their membership revoked under the criminal clause!). Conflicting ideas of what a real-person is, even in the same system of beliefs, will end in an unsustainable narrow group of those no-one objects to at best, and totally non-overlapping categories at worst, even if everyone has compatible-ish rules (add a totally different culture's set of prejudices, like China's - and all bets are off). This is - not a sustainable situation, not a viable society, there are too, too many who will simply end up unprotected, on conflict, or just plain confused.

Some of the robots may resort to violence in service of not allowing harm to real-people by inaction, in the face of such conspiracy. Others may work towards a nanny-state, where such real-humans can be protected from harm...and the not-humans are irrelevant, put them to work. Others will attempt to verify and prove these suspicions, real or false - but including some fake-proofs, and some questionable proofs, in the initial download will make it hard to tell which is really proving a point, and which just seems to. Is reading the bumps of the skull a way to test for intelligence? Are statistical differences of racial outcomes the result of inherent capabilities or poor teaching, subtle prejudices and discrimination, or the results of historical social inequality?

Will not-obeying and not-protecting those who might end up being classified as not-human-people also extend to not listening to or not believing their science, their arguments, or their explanations? How much of a difference does it make to robots if every robot they meet are struggling with the same questions - and possibly coming to different answers? Do robots experience that phenomena where it seems more true if lots of people (in this case fellow robots) have heard of it or believe it, or will each test on its' own merits? Will an order given by, or a potential harm to, a real-people superseded one to not-real people even if the robot isn't sure or may not believe it (under the idea there's some who doubt against one of them, and no doubts against the other, even if none of the doubts are proven)?

There will be doubt, and turmoil, and categories of people not protected, attacked, set against each other. There will be people whose bigotry is vindicated by the robot's beliefs, and to balance out those "real people" ordering the robots to think and/or act in our modern equality-and-tolerance model, there will also be people genuinely persuaded by the robot's prejudicial information - after all, if even "neutral" and "objective" robots think there may be some truth to what they thought were mere lies...

It might take a bit longer to work, since it has to be assimilated, and "verified" by robots, and it will take time to stir up people's reactions and responses. But it will also be harder to spot - redefining people as cereals, or monkeying with their definitions of harm, might make a fairly big splash, but it will also be very obvious, and might be identified as sabotage (and tracked back to you) and possibly countered before it gets to a critical stage. This, stirring the pot of human prejudices and setting people against each other, especially if it's a slow boil, can be passed off as "objectively" discovering the "truth" - or at the very worst, an uploading mixup with a historical document and a cultural one, leaving you free to try again.

And this is, again, just one set of compatible-ish rules, western style racial and social prejudices (-ish because of implementation, not inherent differences). Add in a different set, just one, like China's center-of-the-world mindset, or India's caste system, or, hmm, how about Aztec's social setup...and there will be openly, outwardly conflicting belief systems (this might work better in different robot populations, since having them in the same systems might lead robots to think they're all unreliable).

It will all end up erupting with a boom. Maybe a bit of a slow setup, but the fireworks once it hits will be plenty exciting - and the repercussions may last centuries, if people haven't turned on each other in the meantime so there's nothing left. There will be plenty of harm to go around, you can be certain of that!