[Kind of a rambling collection of thoughts that I tried to organize a bit. Hopefully it's interesting to read]


I've recently had several conversations in which a member of the rationalist community made the case that exploiting people from less economically developed countries for profit is acceptable. In the first, the exploitation was for medical testing, in the second, factories (actually, sweatshops, but that seems extra-loaded, and I'm already using the word "exploitation") were used as the example. I'm not a very eloquent speaker and I tend to get myself tied in knots trying to explain what I'm thinking, then get frustrated that I can't verbalize. It's not a recipe for a productive conversation, so I thought I'd try to write my thoughts down rather than speak them.

Before I continue, I intend the word "exploitation" to mean: Using someone else for your own benefit in a way that harms them to some degree. This doesn't mean that they are harmed more than they benefit (assuming you could compare harms somehow), it just means that they are harmed in ways that they otherwise wouldn't be. The more harm, the more exploitation. There is something of a spectrum of exploitation here- I'm exploited by my employer for profit (not very), a soldier is exploited by their country for violence (a bit more), all the way to wildly imbalanced and unfair relationships like slavery.

The actual definition says something like "using someone for your own gain unfairly" but has a second definition for exploiting resources. Those feel like the same usage with varying levels of harm- if you exploit someone unfairly, then you are causing more harm than benefit. If you "exploit" some natural resource, then you're not harming anyone (directly). So I think my usage isn't entirely unreasonable. If you disagree, feel free to mentally substitute some other word.


The argument that I've heard twice now goes something like this- assume I have capital and I'm trying to make a profit, advance medical research, or accomplish some other goal. I could hire people in a richer country, pay them more, and be able to make less profit, or I could shop around and find a poor country where I could hire more people for less money and make a larger profit.

So I build my factory or medical research facility in the poor country and each dollar spent on wages will produce more widgets (or medical data).

I am a large employer, and by moving my operation to the poor country, I'm providing jobs which give locals more options for work (and presumably more bargaining power with other local employers) at the very least, but I may also be paying slightly more than other businesses in order to attract employees quickly and reduce down time in my business due to turnover. My business, on average, is more beneficial than harmful.

We may harm the locals through long working hours, out-competing local businesses, accidents, poor working conditions, or side effects from medical testing, but we also provide employment choice, slightly increased pay, and purchase some of our materials locally, driving local economic activity. You can tell that we provide a net benefit because the locals willingly work for us- we're not forcing anyone to do anything and it seems plausible that if the harm out-weighed the benefit, they would refuse to participate and I would have to raise wages or take my business elsewhere.

To summarize-

  • We build a factory in a poor country with cheaper labor (meaning more profit)
  • The factory causes both harm and benefit to the locals
  • We can assume it is more beneficial than harmful because-
  • We raise wages (increase the benefit) until people willingly work for us

My disagreement comes in two parts:

  1. Assumptions are made in this example that aren't being explicitly acknowledged, and I think they weaken the argument for exploitation.

  2. In practice, shopping for cheap labor sets up incentives and power disparities that lead to very bad outcomes.

No one that I've discussed this with seems interested in talking about those bad outcomes. They would rather talk about the toy example, rather than how it will (obviously, in my opinion) immediately lead to widespread suffering when introduced to the real world. I admit, again, that I agree with the simplified example and I'm not arguing for any particular policy- I don't know if trying to institute bans or laws against the practice would work. I don't know what the appropriate action should be.

But it's deceptive to argue for the model and then use that as anything other than very weak support for the actual implementation.



The scenario assumes, or seems to assume, a few things-

First, it assumes that the locals are making a free choice to participate in your business, and are therefore responsible for any harm that they receive, while you, as the provider of the opportunity, are only responsible for the benefits you introduce.

In a wealthy country with several hundred years of economic development, hundreds of thousands of job opportunities filling every conceivable economic niche, where the average person is at least minimally educated and somewhat secure from violence- this is a reasonable assumption. It's still technically wrong, because you could argue that everyone has to work for someone or perform some action in order to not starve (at a minimum, they have to acquire or convince someone else to acquire food and insert it into their body), but it's accurate enough.

In a country plagued by internal and external conflict, political strife, food insecurity etc, is it really a free choice? If the alternatives are subsistence farming or working in local businesses that could disappear tomorrow because they don't enjoy support from rich foreigners, then it doesn't seem like much of a choice at all. It feels a lot like shopping around for people dying of preventable diseases and then offering a temporary cure in return for labor.

Second, it assumes that because people will work for you, you are more beneficial than harmful. This is not an assumption that you can make unless your workers are spherical and operate in a vacuum. In reality, people willingly do lots of things that are either harmful to themselves or useful in the short term but harmful in the long term.

This doesn't mean that we should try to pass laws to stop people from eating fast food or watching more than ten consecutive hours of television, but it does mean that we can't rely on them to be perfect signals when we're trying to judge whether a plan causes suffering or not.

Third, this model assumes that a little bit of help is better than none at all. I don't believe that this is always the case. As an unrelated example- I've met a number of underpaid software developers who were self taught and working at large software companies. When they entered the workforce, they had the choice of either pursuing work immediately with what they knew, or living in relative poverty for a period of time and trying to get a degree. They took the first choice and are now in a position in which they can't easily quit their job and go back to school- they have families, bills, and responsibilities that make it very difficult.

They're trapped in a local maxima and the only way to get out of that trap is to temporarily lose a large chunk of their income in order to get a much larger income later.

Given time, a poor country could develop businesses that provide work and opportunities to locals but that also keep the produced wealth inside the country rather than siphoning it off to the already wealthy home country of the founders.

Fourth, it assumes that because market forces are "natural", then they must be right in some sense. Maybe "inevitable" is a better word to use here. This was more of a "vibe" that I was getting from the last person that I had this discussion with. He seemed somewhat baffled that I would argue against the market- Obviously corporations are going to move their factories to areas where they can take advantage of cheap local labor. It's inevitable. Trying to force them to conform to some squishy, vague system of ethics when they have a profit to make is irrational and pointless.

I'm not going to spend a lot of time arguing this assumption because it's likely that no one would admit that they made it. It was more of a feeling that I had during the discussion that I couldn't put my finger on until much later. If I'm not imagining this, and the assumption was made, then I would like to gesture broadly at modern medicine, space and air travel, automobiles, computers, and air conditioning.

Human beings find ways to avoid the natural order all the time- why should market forces be exempt?


We create incentives when we introduce wealthy outsiders to a local economy- outsiders who are invested (both literally and figuratively) in keeping the locals poor and the local economy in shambles. As long as the area around the factory (which, remember, is only there for the cheap labor) is destitute, the factory makes higher profits than it otherwise would. As soon as the local economy begins to compete its way toward higher wages, the value of the location as a destination for factories begins to vanish.

There is an incentive here to keep the locals poor and reliant on the jobs provided by the factory. Because of the wealth disparity between the foreign company and the local economy, there is also a power disparity that provides the means.

During the same conversation in which the factory scenario came up, the group also discussed taxes. One member of the group who isn't originally from the US asked, confused, why we had to do our own taxes if the government already knows what we owe (or could, with minor changes to the tax code). The answer is because there is a tax preparation lobby that benefits from a complex tax code and government inaction.

Imagine this same sort of regulatory control, except the companies are almost as wealthy as the government itself, and are actively trying to sabotage any increases in pay or economic mobility rather than just make a few bucks once a year. It's a recipe for a humanitarian disaster. And a large corporation influencing the local government to create/keep policies that negatively impact the local population is only one way that a properly incentivized corporation could cause harm. Other ways include-

  • Buying and then closing other businesses that compete for the same workforce
  • Purchasing and holding land that could be used by rival companies
  • Dumping waste that causes sickness (in combination with lobbying to avoid safety regulations)
  • Purchasing housing used by workers in order to reduce overhead and limit worker mobility (raise rents to reduce worker turnover)

Or you could just convince the local government to kill your workers when they go on strike:

In Colombia, on Dec. 6, 1928, Chiquita—then the United Fruit Company (UFC)—got the police and army to massacre hundreds of banana workers striking for better conditions. Colombians still refer to the so-called "masacre de las bananeras."

The point is not to argue that any of these things absolutely will happen, it's to point out that there are brutally hard limitations on the real-world application of a simple model like "manufacture widget X in country Y, raise wages until country Y citizens will work for us." Especially when the organization doing the manufacturing has selected the country specifically for it's poverty and lack of better options.

I'm not a historian or an economist, but I don't think that it takes one to know that introducing [organization that exists to make money] into [environment in which making money most efficiently causes mass suffering] is a recipe for mass suffering.


So, to summarize, I think the simplified example sidesteps a number of assumptions. Those assumptions make it less-obvious that shopping for cheap labor is beneficial more than harmful to the locals. In addition, it sets up one hell of an incentive to keep the local region in the same shape as it was when the factory was created- poor and powerless. On top of that, by virtue of the same features that drew the factory builder to the location in the first place, there is a wealth disparity between the intruding corporation and the local economy that gives the corporation the leverage to act on that incentive.

If you're arguing that building factories in poor countries is cost-effective, then I think you can dismiss this entire essay- it doesn't matter whether it's slightly more helpful than harmful. But each time I've heard the argument, it's that shopping for cheap workers is both good business sense and good for the people that you're using for labor, and I don't believe that that is the case.

New Comment
8 comments, sorted by Click to highlight new comments since:

In a country plagued by internal and external conflict, political strife, food insecurity etc, is it really a free choice? If the alternatives are subsistence farming or working in local businesses that could disappear tomorrow because they don't enjoy support from rich foreigners, then it doesn't seem like much of a choice at all.

If the other alternative are that bad, giving a new alternative increases the amount of choices is benefitial. Providing the workers with food security likely matters to them.

As long as the area around the factory (which, remember, is only there for the cheap labor) is destitute, the factory makes higher profits than it otherwise would. As soon as the local economy begins to compete its way toward higher wages, the value of the location as a destination for factories begins to vanish.

If cheap labor is all the people who create sweatshops would care about we would have a lot more sweatshops in Africa. Sweatshops care about factors like infrastructure, political stability and often also the talent pool. 

Whether or not sweatshops generally take actions that damage the surrouding opportunities for work to be able to compete better is an empiric question. There's a high incentive to tell stories like that in Western media and for Western academics to run studies that find those effects. 

There are many ways in which you can run a sweatshop that produce damage to the local economy but if we look at the decision to build a sweatshop we can decide to build one that's benefitial to the local economy. 

The practical question is whether or not the effort that current multinationals exert on sweatshops to operate in an ethical matter results in sweatshops that are worthwhile to have. 

Is this a specific case of the general argument that "you can't ethically financially interact with people who are sufficiently poorer than you(r reference class), except through charity"? I think the arguments here:

  • Sufficiently poorer people might not feel able to reject your offers
  • SPP might not be able to make the best decisions
  • You might be motivated to keep SPP that way, if you have the option of getting any utility from trading with them

are more general arguments than the specific case of factory building. I don't have a good answer to them, but kinda feel that it's disrespecting SPP's agency somehow?

I do not have a full solution to this, but my intuition says this: Suppose that it is ethical to employ your neighbor under certain conditions, and that you are already doing that. Then you find out that the average wage in Pooristan is 10 times lower than in your country. Ignoring the effect on your neighbor who may become unemployed when his job is outsourced, I would probably say that:

  • It is morally neutral to hire a person from Pooristan to work under exactly the same conditions, only for 10% of the salary you would offer in your country.
  • The right thing to do is to hire a person from Pooristan to work under exactly the same conditions for maybe 20% of your local salary. That way, you share the profit from the international trade. Note that you still profit a lot from this arrangement.
  • It is an exploitation to leverage this situation into offering worse working conditions, such as longer hours, health-damaging workplace, or abusing your employees in some other way.

[ I'm in favor of discussing this, but I'm not sure it's framed compatibly with LessWrong normal assumptions, and I'd rather not see a lot of it here.  As such, I've neither upvoted nor downvoted. ]

Leave aside that your definition of "exploitation" covers the majority of human interaction - there's always a mix of benefits and costs which are different (and on different dimensions) to each participant.  The discussion can proceed even with that bias.

My primary concern is that you're leaving out the MOST IMPORTANT part of decision-valuation: compared to what?  Taking lower profits rather than killing striking workers is a comparison that I think most of us understand (and prefer the first option).  Building a factory or not leaves out an analysis of "not", in a way that it's impossible to answer which option is better.  Building a factory in a more industrialized/expensive location is one option to compare.  It's lower-profit, but it doesn't help (and possibly further harms, through loss of opportunity) the poor area that remains subsistence-level in output.  It probably does (maybe) help the rich-enough-to-make-good-decisions area, but it's a less important change for them than it would be for the poor area.  

Extremely great to see this type of discussion show up here! I've been looking forward to folks posting more stuff in this category.

The usual argument I see is "well, yes, you are absolutely stealing lifeblood from the people you exploit, but it's worth it to end the horrors of ancient earth faster". My question for people who agree with that statement is, why are you confident that exploiting the market vulnerability of these folks being underpaid trades off sufficiently to warrant not participating in the FDT group of "people who only make trades that cause all parties to reach baseline human non-suffering"? it seems to me that this "worth it" argument is made of a kind of failure that typifies an unsafe ai: there's a system-local causal impact on people of the exploiting org; while the vulnerability of the exploited market does mean it's possible to not fix the negative parts of this causal impact, and while there may be the possibility of turning this into aid for more people later, wouldn't it be worthwhile to move the needle on ratio of gains-from-trade towards giving enough of the gains to the folks you hire that you cause them to reach baseline non-suffering human experience?

I don't agree with you but I like that you're exploring this.

Rather than try to go through a bunch of points, let me just pull one thread about "free choice".

You argue:

If the alternatives are subsistence farming or working in local businesses that could disappear tomorrow because they don't enjoy support from rich foreigners, then it doesn't seem like much of a choice at all. It feels a lot like shopping around for people dying of preventable diseases and then offering a temporary cure in return for labor.

There's two way I'll approach this. First, I'll assume that some notion of "free" choice is coherent.

But this is a choice freely made. Was it made at gunpoint? Perhaps sometimes in history but in this case not. This is people following the incentives. If you think the incentives that lead to this are bad, then this quickly unravels to thinking all the incentives that lead to anything beyond the state of living in small bands of hunter-gatherers is not a free choice because in each step along the civilizational "ladder" we give up things people found valuable about their old lives to get other things. Some people do think we made the wrong choice, but most people are happy to live with our nicer material conditions in exchange for having to deal with the psychological and meaning issues created by living in agricultural, industrial, and service economies.

Basically I think you're making an argument that living in modern society is also not a free choice. Which, if you are, sure, but lots of people disagree on the reasonable grounds that they like the tradeoff of material comforts for psychic distress.

(There's also something where I think your arguments take away the agency of these local people and replace it with paternalism about what you think is best for them, but that's a separate concern.)

Okay, now to disagree with the very notion of "free" choice.

What would it mean to make a choice "freely". Do we have free will? If not, in each moment we act in only the way we ever could have acted, so in some absolute sense there can be no exploitation because everything is always going to happen just as it would. Free choice and free will appear to exist in a relative sense, and it may be useful to think in an ontology with free will to get desired outcomes (again, from within the relative framing), but that doesn't mean they are anything more an illusions that help us make sense of the world as we find it rather than something inherent in the world.

Yes, this sort of a fully general argument against concepts and words, but I bring it up because you're trying to be precise about what free choice is, and I think that's not really possible. Free choice is something intersubjective and exists along the dimensions of what feels free (more likely, fair) to us, and you'll have to convince folks not that the choice is not made freely, but that their notion of what is free and fair is mistaken (which you do try to do elsewhere).

I appreciate your reply- it was thoughtful and lucid. Thanks for taking the time to comment. Well, I thought I wasn't going to have time for a long(-ish) reply, but once I started writing I couldn't stop, so here you go-

First, I'd say that "free choice" obviously isn't an absolute state. But I think that there are more or less free choices. Working for a factory in the US is a more free choice while working for a factory in a country that has no other good options is a less free choice. I could have been more clear about that.

Second, I'm not going to argue for or against free will itself, because whether or not it exists, we have the appearance of free will and have to make decisions as if it does exist. It's similar to the simulation arguments in that way- maybe we live in a simulation, but it doesn't really matter from a practical perspective.

Lastly, I don't believe that it's paternalistic to evaluate the possible outcomes of my own actions (as the factory owner) and change my behavior based on them in order to conform to my own system of ethics.

It's funny, but during the original conversation that led to this essay, the person arguing for factories was arguing that it's the best way to pull poor countries out of poverty. He was accused of paternalism as well, so that argument goes both ways.

Thanks again.

I strongly disagree with this view; "freeness" of a choice should be analyzed in terms of how much a local system's thermal noise has causal impact on other systems. For example, if a system such as a brain has free agency over an object, such as a rock, then the noise in possible brain activations has determining impact over what trajectories the rock will follow, because the denoising process the brain implements gets to guide the behavior of the rock. Notably, this means that even with a deterministic PRNG, diffusion models have a very significant amount of "free agency", because they have full control over the trajectory and ability to reject noise. Other generative models do as well, but diffusion has notably more than others due to how many steps of denoising steering the model takes; the model becomes very strong at rejecting noise and making its choices, and so the model will always have impact on its output that controls it into aesthetic states defined by the model weights.

Compare this to how much aesthetic control a brain has over their body's behavior in circumstances with reduced optionality; the reduced optionality is recognizeable from an oracle's perspective due to the fact that the externally unpredictable chaos of local trajectories in the brain does not have as many paths that cause the brain's aesthetically desired outcomes.

For example, if someone is the only person on a planet, the person cannot push through barriers in statespace such as "I'm low on fuel and have an unknown disease".

Or if someone is in a place where the people around them consider them to not have offered acceptable trade for access to material goods and services, but the person is out of fuel and has an unknown disease, their optionality is significantly reduced versus if material goods had not been collected by the actions of the stronger agents in their environment.

So, okay, I've claimed we can define optionality usefully. What does this mean about how a stronger agent should behave in a competitive environment?

I'd argue that the key question any safety-of-inter-agent-interaction alignment approach needs to answer is how to distribute gains from trade in order to maximize optionality of both agents in the trade. In other words, to minimize externality-inducing exploitation of vulnerability, maximize the combined optionality of both agents.

I tried to find papers to cite for this post, and didn't find any of what I was looking for, which is frustrating, because I'm pretty sure I've seen good work out of economic game theory recently, but my searches are coming up mostly empty handed. here's one result: