Jonathan Zittrain, Professor of Law and Faculty Co-director (and co-founder) of the Berkman Center for Internet and Society at Harvard University, gave a presentation at the Computer History Museum in Mountain View about a month ago that ought to be required viewing for anyone interested in Cloudlabor and Crowdsourcing.
Drawing examples from all over the Internet – including a certain iPhone app that you may have heard of – Zittrain raises some serious (and some seriously entertaining) questions about ethical and legal aspects of distributed human computing.
Straight from the Berkman Center YouTube channel, here’s the full video (which is also available for download under a Creative Commons Attribution 3.0 license from the President and Fellows of Harvard College:
Zittrain focuses on the potential alienation and opportunities for abuse that can arise with the growth of distributed online production. He also contemplates the thin line that separates exploitation from volunteering in the context of online communities and collaboration.
I enjoyed his analysis and the discussion afterwards, although I suspect that some of the conversation with the audience might get lost in the video. As with Zittrain’s most recent book, The Future of the Internet and How to Stop It, this is some of the best thinking about life online that you’ll find anywhere.
Zittrain has also published an abbreviated portion of his argument in Newsweek under the slightly more extreme title “Work the New Digital Sweatshops.”
I find a lot of what Zittrain has to say compelling; however, I do wonder if the efforts of ReCaptcha-spammers and sock-puppeteers to exploit Crowdsourcing markets will ultimately prove successful. I also wonder whether the imposition of labor regulations in these contexts makes sense or would prove effective. Should my decision to kill time or make a few extra bucks by filtering images be subject to labor law? What about the ability of other people to offer money for distasteful and perhaps unethical (but usually not illegal) micro-tasks?
It may be a few years before anyone really understands if Crowdsourcing lends itself to unique types of market failure along these lines, but Zittrain and others such as Lily Irani and Aaron Koblin are doing us all a favor by asking some of the most important questions early in the game.
Full disclosure: the author of this post is affiliated with Harvard and the Berkman Center for Internet and Society, where he was a fellow during 2008-2009. While he doesn’t think that his affiliation influences his opinions about Zittrain’s work, it does mean that he’s very pleased not to be spending another winter in Cambridge this year.
Ryan Shaw
12/22/09
I also wonder whether the imposition of labor regulations in these contexts makes sense or would prove effective. Should my decision to kill time or make a few extra bucks by filtering images be subject to labor law?
Your questions echo those of U.S. business leaders a century ago, who also extolled the virtues of cheap labor and fought anti-sweatshop regulation. Maybe you should read up on the history of the labor movement before deciding whether you want to follow in their footsteps.
aaron
12/22/09
Hmmm, that doesn’t feel like a fair characterization of my statements (or my knowledge of U.S. labor history), Ryan.
Obviously, you’re welcome to make the claim that Crowdsourcing and other forms of distributed collaboration online are analogous to early 20th century factories, but that seems like an over-simplification in many ways.
Indeed, most of the survey data that I am aware of suggests that participants in distributed labor markets and online communities that produce user-generated content tend to be quite privileged (for a great example, check out Eszter Hargittai and Gina Walejko’s paper, “The Participation Divide” ). If nothing else, I think that complicates the analogy to sweatshops, where workers tend to be less wealthy and the conditions of exploitation more extreme.
Ultimately, I think there’s a lot of gray area and limited reliable evidence when it comes to theorizing the relations of production and power in the sort of online settings that concern Zittrain. I agree with you that industrial factories and labor relations one hundred years ago provide useful examples for comparison, but such a comparison is messy at best and may not provide an ideal basis for designing contemporary governance institutions.
My intention was not to dismiss the importance of thinking about exploitation and regulation online, but to underscore some of the difficulties involved. Thanks for the provocative comment and apologies if you (or anyone else) finds what I said to be offensive.
John
12/22/09
I’m going have to side with Ryan here. There was that fire in the turker chat room last year that killed all those people–apparently Amazon had chained the doors shut. Plus three people fell ill just last week when they got emailed some bad spam; the FDA traced it back to unsanitary conditions on Amazon’s servers (someone was solving captchas without gloves on). This might just be a rumor, but I heard that when Jeff Bezos was touring the turker factory floor, some poor guy got his arm ripped off in a hadoop.
Ryan Shaw
12/22/09
Aaron, I’m not offended by what you had to say. But I’m glad I provoked you a bit. I was disappointed that your reaction to Zittrain’s talk didn’t go further than posing a few hypotheticals and then reverting to the default Silicon Valley “regulation == bad” setting. So I wanted to push the point a little further.
It may very well be that current participants in distributed labor markets are privileged and doing it to “kill time” or “make a few extra bucks.” But what is it about the structure of distributed labor markets that ensures that this will continue to be the case? Certainly there are efforts to spread crowdsourcing in developing countries. Do we have to wait until there are documented cases of crowdsourcing sweatshops before considering how to update labor laws?
John, labor law is about a lot more than ensuring a safe and sanitary workplace. It’s also about making a living wage without having to work 20 hours a day, having opportunities to build skills and advance, and the right to organize. Mechanical Turk doesn’t have built-in mechanisms for ensuring these rights.
I think a lot of the uses of crowdsourcing are very cool. But we techies have a bad history of focusing on the coolness and pushing off consideration of the consequences until very late in the game.
John
12/22/09
Consider a particular Turker whose best option outside of turking pays $2. When they work on AMT, they can make $3 and suppose the value of their work to a requester is $4. Both the requester and the worker are getting a $1 surplus.
Now suppose the Click Worker’s Local 327 decides that a “living wage” is $5 day and AMT complies. In this scenario, all the buyers exit (they will not pay a penny over $4, obviously), and the turker goes from making $3 to $2. They lose a $1. Buyers lose a $1. Who is made better off?
I think your implicit assumption is that requesters actually get much more surplus and that one could impose a price floor and transfer some of that surplus to workers without affecting the demand for labor (this is perfectly consistent with your comment about this history of the labor movement). I grant that this might be possible, but I’m pretty skeptical that anyone knows or could learn enough about the private valuations of buyers and sellers to improve the welfare of the market participants through crude interventions like price floors.
Ryan Shaw
12/22/09
John, I’ll just point out that your “helping them will hurt them” argument is precisely the kind of argument that historically has been made against any kind of labor standards. See Richard Rothstein’s article comparing the arguments industrialists and academics made against child labor laws then and the ones they make now.
In any case, I wasn’t arguing for a price floor. I’m arguing for the engineers and economists who design crowdsourcing markets to consider the welfare of workers as carefully as they do the satisfaction of requesters. Surely they can come up with something more clever than price floors. And if not, perhaps the threat of regulation will motivate them to do so.
John
12/23/09
It would really be better to address the argument I make than to say it looks like some other, presumably false, argument. We aren’t talking about child labor or activities that poison the planet—we are talking about a very particular kind of safe, convenient work that’s available to people with very few outside options.
I reject the claim that online labor markets like AMT, oDesk, Elance etc. are designed for the benefit of buyers—they are designed for the purposes of their profit-maximizing owners. Think for a moment about how all of these sites make money—they take a fraction of all transactions. Unlike requesters (who do want to see wages fall as low as possible), the creators of these markets have strong and direct financial incentive to maximize the total wage bill. For any given volume of work, they’d love to see prices rise. I’ll agree that no private company will consistently and voluntarily act for the public good—but I have a hard time accepting that they are stupid or act against their own financial interests by ignoring the essentially two-sided nature of their business.
If you aren’t proposing a price floor, fine, but that’s what a living wage is. If you have some other realistic policy that would enhance the welfare of workers without entailing a large risk of destroying said market, we’re all ears. Maybe there is something that could be done—I like Zittrain’s idea of letting workers make verifiable profiles of their work history when they change markets—but threatening to put firms out of business with intentionally destructive regulations is a pretty poor way to develop policy.
Ryan Shaw
12/23/09
My point about your argument was that it’s easy to construct toy models that prove that the arrangement you favor is the best possible one. Such models conveniently exclude all the aspects of reality that make crowdsourcing potentially problematic. For instance: we aren’t talking about child labor? How do we know that? I see nothing to prevent someone from setting up a warehouse in Cambodia with a T1-line, a bunch of cheap Dells and a few dozen 14-year-old clickworkers. Does Amazon send around monitors to ensure this kind of thing doesn’t happen? I don’t believe this is actually happening much now outside of some MMORPG goldfarmers, but if crowdsourcing becomes as economically important as its evangelists claim, I see no reason why it won’t.
You keep stressing safety, but again, safety is not the only concern. Many people oppose garment industry labor practices in developing countries, not on the grounds that the factories are unsafe, but on the grounds that workers are paid extremely little, have no opportunities for advancement, and are subject to termination for attempting to organize. It’s true that the garment factories present the best option available for many of these workers, but at the same time they also ensure that no better options can arise. Can crowdsourcing claim to be a superior arrangement?
You may be right that the operators of online labor markets try to balance the interests of requesters and workers. The way they market themselves seems to me to be heavily slanted towards the former, but perhaps that’s not a good way to evaluate them. My concern is that they will find it easier to expand their reach to larger, less-privlleged populations–thereby increasing profits through increasing volume–than to invest in retaining higher-paid workers.
I agree that Zittrain’s idea is a good one. I also like Lilly Irani’s idea of profiles for requesters, so that workers can make more informed decisions about who they work for. In general, I’d be in favor of any policies that enable workers to be better informed about what it is they are doing, who they are doing it for, and who they’re working with or competing against in the labor market. I suspect that you will consider these “unrealistic” policies that will “destroy” the labor-as-a-service business. But if that business only works because it removes safeguards that labor activists worked for decades to enact, why shouldn’t it be destroyed? And if adding those safeguards won’t destroy it, then why not enact them?
John
12/23/09
I should first say that I’m enjoying this conversation – thanks for replying. I think this is an important issue and I don’t think there’s been enough thoughtful back-and-forth.
So let’s return to Cambodia example, because I think it’s a good one. Assume that we pass a law that requires AMT to send a representative to verify the identity of all account holders. What’s the minimum amount this would cost AMT? Even if they could use local affiliates, my guess is that the cost would exceed AMT’s expected per-worker profit and the market would shut down. I know this looks like a catch-all anti-regulation argument, but it’s not. I’m suggesting that we need to look at each piece of proposed policy on it’s own merits. Because the transactions in online markets are for so little money and the products have fairly low value, the imposition of cost-raising regulation has to be pretty light.
So let’s look at the proposal to require AMT to verify worker identities with a periodic, in-person visit more closely. What’s the expected benefit in terms of human welfare for ensuring that no children are working online? Even if it does not shut down the market, the benefit might be far lower than we think. For example, developed country legislation designed to limit child labor in developing countries doesn’t magically put displaced child laborers into high-quality schools and enroll them in generous public assistance programs—it forces them into more dangerous lines of work like junk scavenging and prostitution. These people face a horrible trade-offs; in my opinion, online work is one of the least bad options.
As far as the low wage critique, I think my earlier simple labor supply model is still valid. I don’t deny that it is possible to raise wages by fiat (if not through a price floor, then through regulations that lower the supply of workers or increase the demand for their services) without reducing the volume of trade, but I don’t think it’s likely. I think it’s much more likely to curtail the volume of trade. Higher prices would be great for the people who have the skills to get the now-higher prices work, but it would likely price the poorest, lowest-skilled people out of the market and make them worse-off.
I don’t think the argument that worker’s cannot advance has much bite in more important online labor markets. Look at oDesk—you can see remarkable wage growth and plenty of examples of worker’s acquiring new skills and earning wages 3 or 4 times higher the averages in their home country. If we required oDesk to verify the birth certificates of all buyers, it would likely kill the market and all the positive benefits these developed world / developing world labor market exchanges provide.
Now – to back up, some regulations do make sense because they impose little costs but improve welfare. Zittrain’s idea falls into this category. Turkopticon is brilliant. There’s surely more like this that can be done and I think it makes sense to focus on solutions that can expand the extent of these markets, help improve skills, reduce the potential for abuse. However, the most important moral fact is how much the world’s poor could benefit from access to first world labor markets and how much damage we could do with ill-advised regulation that destroys markets.
Lilly Irani
12/26/09
I wonder if the UFC and TV comment spam was generated through AMT work. There are a decent number of tasks that tell people to go comment on blogs.
It’s been great to follow the debate here.
The only thing I would add is that first-world markets as good for the world’s poor is a huge assumption and probably at too general to be helpful. There are a lot of different situations among “the world’s poor.” Some are poor because they lost their arable land to white settlers 50 years ago — is getting them to labor for european industry the best way we can think of to help them instead of, say, figuring out how to repatriate or redistribute some of that arable land? (See Ferguson’s “Anti-Politics Machine” for a good treatment of this situation.) Others are poor because they are refugees from a war perhaps stemming from conflicts over water between tribal groups that didn’t exist in strong forms before european colonial rule. Others might not even think they’re that poor and have to be convinced that they are so by the NGOs who are in the business of development (See Shepherd, “Agricultural Hybridity and the ‘Pathology’ of Traditional Ways: The Translation of Desire and Need in Postcolonial Development,” 2004). So when we talk about the poor, we should recognize that first, there is no “the world’s poor.” There’s different issues of livelihood in different places because of different causes. Second, in many such cases, a first-world market job might be a bandaid on a much larger wound. And third, employers and third-parties who “give work not aid” should own up to their interests in providing a bandaid that also allows them to reap profits by securing inexpensive labor.
One of the interesting and tricky problems with AMT is that it spans so many countires, legal jurisdictions, and cultural situations. Working out of Scandanavia while you’re also on welfare and have a public health plan seems really different than working somewhere where this is the only source of income you have, there is an incentive to work hard short term, and you have no health plan or safety net should long term issues like RSI happen. AMT labor conditions are different in different places.
Alek Felstiner
12/27/09
I was going to say the same thing as Lilly – as someone who’s read a few HITs, those look a lot like a $0.03 task.
I tend to agree with Ryan, insofar as I consider the analogy to earlier paradigm shifts apt. Our current raft of regulations – protecting individual rights, safeguarding collective action, ensuring health and safety, prohibiting discrimination, etc. – arose out of a wholesale, revolutionary shift towards industrial production that completely upended prior notions about the employment relationship, bargaining power, contractual relations, and the (necessary) role of government in regulating employment through the commerce clause.
Those New Deal and mid-century laws do _not_ reflect the subsequent shift to service economy that took place towards the end of the twentieth century, and they have very little relevance in the growing information economy. So it’s good to question their applicability, but that’s different from deciding that because our current structure does not appear to contemplate (or in some cases, even apply) to crowd labor, regulation will be automatically inappropriate.
Also, a weakness in both John’s and Ryan’s arguments is the presumption that an entirely new mode of production (as opposed to a variation on an existing theme) would be governed by the same old economic doctrines. That isn’t to say that supply, demand, scarcity, etc. have no meaning in the crowd labor context. But trying to evaluate opportunity costs and transaction value in this frontier is a fool’s errand.
I think Zittrain is right that, instead, we have to figure out what we can and cannot accept, in terms of dignitary interests, proprietary rights, etc., and work from there. I also think that one of the best ways to find an appropriate balance is to test the waters – we have employment law in this country, and we should put it to use. It runs against bedrock legal principles to allow a single crowdsourcing platform, like Amazon, or a coalition of them, to decide that what they do should fall outside the ambit our our existing employment law regime. Same goes for commentators.
That’s one thing that bothers me about this and similar debates, as a lawyer. Absent legislative action, it’s up to the courts to decide whether the Fair Labor Standards Act (fed’l minimum wage & overtime law) applies to crowd workers. And the decision has very little to do with economic arguments – that’s not how the statute (or similar state statutes) is structured. The same goes for the other laws that regulate the workplace. So we can certainly debate whether a minimum wage, for example, would be _good_ for mTurk, but that’s an entirely irrelevant question from a legal perspective. What matters is a) whether crowd work crowd employers actually fall into the existing regulated categories, and b) whether – given the rationale behind those existing categories – crowdsourcing should be regulated.
Finally, it might be worth noting that Amazon et al are actually way ahead of us. They’ve anticipated all these questions, and have attempted to conform to existing laws or subvert them, in their terms of use. I’m writing on this now, but it’s kind of a hornet’s nest of legal jargon – wait for the upcoming paper on SSRN, I guess.
John
01/04/10
Hi Alex
Thanks for joining in. I think lawyers have something to contribute, but I don’t think that the economic or sociological analyses of online labor are irrelevant or dispensable because “it’s up to the courts.” If you’re a hard-core legal realist, court decisions might depend on what the judge had for breakfast – hopefully we (e.g., social scientists/theorists/legal scholars etc.) can offer up something better than breakfast as the decisive factor. Even if you are only a squishy legal realist, there’s going to be enough indeterminacy in any statue or relevant body of case law (esp. given how unusual and novel online work is – a point you make) that arguments about public policy will almost certainly be influential (imagine if the first AMT case comes before Judge Posner…).