Class Action Against Crowdsourcing
Something is brewing in the world of digital labour. In October 2012, online worker Christopher Otey filed a class action lawsuit against the US based company CrowdFlower, one of the largest platforms for the completion of so called ‘micro-tasks’. The company claims to have a reserve army of millions of workers and according to its CEO Lukas Biewald, they hire up to 10.000 people per hour and up to 3 years of work per day 1 The pending lawsuit is now challenging the companies failure to pay the minimum wage under the Fair Labor Standards Act to its US workforce and Christopher Otey’s lawyers are searching the web for other underpaid members of the online crowd who want to join the class action. CrowdFlower’s lawyers point out, however, that Christopher Otey did his work completely voluntarily and that he and all the other ‘cloud-workers’ are not employees but free contractors. The case is still open, but it has the potential to shake the foundations of a business model that has been mushrooming around the globe over the last five years.
According to the Crowdsourcing Industry Report, the sector is doubling its workforce and rising its revenues by 75 per cent each year. The revenues of the platform owners that is. The workers earn between $1.40 and $2 per hour, depending on the platform, the job and their experience — the quicker they work, the more they earn — but on average, the income is below the minimum wage of Beijing and a far cry from the US minimum wage of $7.25. As the report also points out: “… more than half of all the crowdsourcing workers live in North America and Europe and workers are generally very well educated. Almost half have a bachelor degree and only 5% are truly low skills workers.” 2
How to come to terms with crowdsourcing? To some it is just a neutral umbrella term describing various forms of distributing labour; to others it stands for the exploitation of cheap or free labour with detrimental effects on workers and professions. But can crowdsourcing be described as exploitative even when all participants are volunteers and know the conditions? Is it still labour when people do it as a hobby? Is crowdsourcing inherently unethical or is it just a question of how the parameters are configured? And how does it effect the design profession?
Buy it, use it, break it, fix it,
trash it, change it, mail – upgrade it,
charge it, point it, zoom it, press it,
snap it, work it, quick – erase it,
write it, cut it, paste it, save it,
load it, check it, quick – rewrite it,
plug it, play it, burn it, rip it,
drag and drop it, zip – unzip it,
lock it, fill it, call it, find it,
view it, code it, jam – unlock it,
surf it, scroll it, pause it, click it,
cross it, crack it, switch – update it,
name it, rate it, tune it, print it,
scan it, send it, fax – rename it,
touch it, bring it, pay it, watch it,
turn it, leave it, stop – format it.
Daft Punk, Technologic, 2005
From the Empowerment of the User to the Harnessing of the Crowd
Every day, we click our way through an endless succession of micro-tasks. Surf it, scroll it, pause it, click it. Isolated, these are almost meaningless but in aggregated form they create data of great value. It is well known that the services of Google, Facebook and the like are not actually free, but payed for with personal data and attention. In other words, the users are the product being sold to advertisers. With the accusation of exploitation already looming, Nicholas Carr has described Facebook’s business model as ‘digital sharecropping’. 3 He refers to Facebook’s average yearly revenue per user, which was at $5,11 in 2011. Not much for a single user, but after all, they got a billion of them. However, as puzzling as it might be that so many users prefer selling out their privacy instead of paying a small maintenance fee, I wouldn’t describe Facebook’s business-model as exploitative. The value creation by the users happens as a side-effect of their activities and in return they get a service that has a high value for them. In the case of Google’s search it is even harder to imagine paying with real cash instead of data and attention.
Write it, cut it, paste it, save it. These tasks are already more demanding because they revolve around the creation of content, be it for self-expression or as a service to others. Amateurs online write articles for Wikipedia, moderate help forums, debug open source software and make valuable contributions to sciences from astronomy to ornithology. With increasing complexity, these tasks stop being micro and demand a high level of engagement and expertise. They eventually become indistinguishable from work. As portmanteaus such as prosuming 4, produsage 5, playbour 6 or pro-am 7 have tried to express the lines between amateur and professional, between play and labour have been continuously blurred. Crowdsourcing harvests the whole spectrum of these hybrid activities and it’s what makes the valuation of appropriate remuneration so tricky. The criticism of amateur work online used to circle around its supposedly low quality 8, but the more relevant question is actually who is entitled to make a profit from all the free contributions.
The concept of user-generated content was central to the so called Web 2.0. The new version of the internet, so it was said, had become more emancipatory and collaborative. But what actually united the new breed of commercial websites that arose from the ashes of the dotcom crash was that they all found ways to let the users produce the content. In the case of Amazon, they already contributed ratings, reviews and recommendations. With the launch of Second Life in 2003 and, most importantly, YouTube in 2005 the concept was elevated to a new level. Now, the users also created the core product. Wikipedia had already started in 2001, but it was between 2004 and 2006 that it was growing exponentially. All this contributed to a great hype about the empowerment of the user, which peaked in 2006 when Time magazine made You the Person of the Year: “The new Web is … a tool for bringing together the small contributions of millions of people and making them matter. … It’s about the many wresting power from the few and helping one another for nothing.” 9 As it turned out, half of this was an illusion. While people do indeed help each other for free, the power seems to be back firmly in the hands of the few. The users had much more control over their data and content on their private home pages, before they handed over everything to the global aggregators and the individual users transformed into the crowd.
The Reinvention of the Crowd
In the nineteenth century, a crowd was still an unruly gathering with a dynamic that could quickly turn a group of cheering spectators into a raging mob. In 1895 Gustave Le Bon established the field of crowd psychology with The Crowd: A Study of the Popular Mind. The sociologist was convinced that the sum of people would always be less then its parts: “Civilisations as yet have only been created and directed by a small intellectual aristocracy, never by crowds. Crowds are only powerful for destruction.” 10 With the spectre of democracy haunting Europe, the question for Le Bon was how to keep the crowd at bay and influence it in favour of those in power; part of his study therefore reads like a manual for crowd manipulation. Occupy Wall Street, the London Riots and the Arab Spring are recent examples that this archetypal crowd still exists. Since the turn of this century, however, the revolutionary crowd in the street has been supplemented by the docile crowd online, productively clicking in the hours. The challenge is not anymore how to suppress its destructive power but how to harness its collective intelligence. This paradigm shift was partly triggered in 2004 by James Surowiecki’s book The Wisdom of Crowds – How the Many are Smarter then the Few. Backed by an array of research from various fields he argued that Le Bon had things exactly backwards: If you put together a large and diverse enough group its decisions will over time be intellectually superior to any isolated expert. 11 And as the success of GNU/Linux and the Wikipedia had shown by now, online communities were not only capable of solving complex tasks, the contributors were also willing to do the job for free.
In 2006, journalist Jeff Howe labeled this new form of labour online for Wired: “Welcome to the age of the crowd, where … distributed labor networks are using the Internet to exploit the spare processing power of millions of human brains. … The labor isn’t always free, but it costs a lot less than paying traditional employees. It’s not outsourcing; it’s crowdsourcing.” 12 Clearly, this was not about the empowerment of the user anymore. Now that aggregation on a massive scale had become possible, experts started thinking about how to put this yet to be exploited resource to good use.
Human Spare Cycles
Attempts to fathom the potential quickly led to astronomic calculations. In Cognitive Surplus: How Technology Makes Consumers Into Collaborators, Clay Shirky estimated that “the world’s educated population has three trillion hours of free time each year.” 13 In contrast, it is said that it took ‘only’ about 100 million hours to create the Wikipedia. What could be achieved if only people would waste less time in front of the TV and devote their free time to more productive causes? Along these lines, the game-designer and author Jane McGonigal has pointed to the total number of hours people played World of Warcraft, which in 2011 accumulated to 5.93 million years. Her solution is to make the world more game like and to create games that tackle real world problems such as health and sustainability. 14 However, this so called ‘gamification’ cuts both ways. Not only makes it games more work like, the introduction of points, badges and other virtual incentives can be very manipulative. It propels competition and ambition among the workers and transforms the feeling of loss of time into a feeling of achievement and progress. 15 It has become a tool in crowdsourcing to ‘pay’ the contributors without having to give them cash, which is why media philosopher Ian Bogost has suggested to better speak of ‘exploitationware’. 16
Luis von Ahn, researcher at Carnegie Mellon University, has developed a very special from of harnessing the ‘cognitive surplus’. He is the inventor of reCAPTCHA, those distorted letters that we have to type in whenever we create a new account online to prove that we are human. When we type in those words we recognise a fragment of a book-scan that a computer could not decipher. Von Ahn says he is treating “human brains as processors in a distributed system, each performing a small part of a massive computation.” 17 The curious thing about Ahn’s method of crowdsourcing is, that people often don’t even realise that they perform ‘human computation’.
The aggregation of usage data, unconsciously performed micro-tasks and even user-generated content are all examples for crowdsourcing in a broader sense. In a narrow sense, as outlined by Howe, it means replacing employees with precarious crowd workers. Criticism not only comes from those threatened by it. In 2007, Jimmy Wales described crowdsourcing as a “vile way of looking at the world. This idea that a good business model is to get the public to do your work for free — that’s just crazy. It disrespects the people. It’s like you’re trying to trick them into doing work for free.” 18 This might come as a surprise from Wales, after all, the Wikipedia was built on free contributions. But the essential difference is that as with open source software, the volunteer work done by the community creates a free resource for the commons. While in crowdsourcing, the creators often have no direct use for their contributions and even hand over their intellectual property rights. The open process ends with a closure and the sole beneficiary is someone outside of the community. The use of the term crowd reflects this outside perspective. A crowd is other people.
Core Methods of Crowdsourcing: Micro-tasking vs. Contests
Crowdsourcing is sometimes used by companies as a one-off marketing stunt to engage customers with a brand and it can also serve as a market research tool. But in its narrow sense, it has become a business in its own right. Specialised companies cultivate permanent communities of online workers, not unlike beehives, and offer their workforce to external clients. The methods for orchestrating the workforce of the hive vary greatly. Some owners install incentives for collaborative behaviour such as ‘Karma points’, others foster competition. Many offer non-monetary ‘gamification’ incentives that give the contributors reputation inside the community, others actually pay their workers. But since the crowd is by definition not restricted in numbers while the sum of money that is being paid for a job certainly is, it is not possible that everybody gets paid in full. When money is involved, there are essentially two different models: Either, the workers receive micro-payment for solving repetitive micro-tasks, for example, they get a cent for each item they categorise. Or there is a contest, or a gamble, in which the workers all do the same job but only one is getting paid eventually. This second model is used when the work is more complex and time consuming and can’t easily be split in tiny units, as it is usually the case with design tasks.
CrowdFlower.com, the company now faced with the class action law suit, is a typical example for the micro-payment model. So is Amazon’s Mechanical Turk, named after the historic chess robot that was operated by a human, hidden inside the machine. Amazon accordingly describes its service as ‘artificial artificial intelligence’ and it addresses the same sort of problems as Luis von Ahn with his human computation, fine grained repetitive tasks that computers aren’t very good at. The most baffling application of micro-tasking is probably Soylent, the ‘word processor with a crowd inside’. It is a plug for text editors that allows authors to assign parts of their writing to the crowd of ‘Mechanical Turkers’ for correction or shortening without even having to leave the programme. 19 Tellingly, the name is derived from the apocalyptic science fiction film Soylent Green. It shows a world suffering from extreme overpopulation in which the popular and nutritious snack Soylent Green turns out to be made of humans — a delicate way to deal with the crowd …
Let Them Design Logos
As it turns out, logo design in particular lends itself to the contests model. There are now dozens of so called ‘logo mills’ such as CrowdSpring.com and DesignCrowd.com. A particular large one, 99designs.com, claims to be ‘the fasted growing design market-place in the world’. It has more than 200,000 registered designers and it already conducted over 180,000 design contests. Even though the site boasts many numbers, the pricing scheme of 99designs is deliberately opaque so it is not directly visible that the site takes a share of 40 to 45 per cent. From the initial $300 that a client is paying for a logo contest, the platform takes off $120 right away. The client gets on average 116 logos, which leaves the designers with a chance of 1 in 116 to eventually getting paid $180. The average renumeration comes down to about $1.50 per logo design, before taxes. There are higher paying contest for things more complex, but the average money paid out per design on 99designs still is only $2. Since designing a logo usually takes significantly longer then an hour, the designs on offer can only be derivative, of low quality or the contributing designers work for even far less then their colleagues toiling away in the micro-payment sweatshops. It is, by the way, the external client that decides who wins and if anyone will get paid at all. 99designs offers a 100% money back guarantee if the client doesn’t like the results.
At first sight, there are some similarities here with the notorious pitches in architecture. As Rem Koolhaas pointed out: “It is a complete drain of intelligence. I don’t know of any other profession that would tolerate this. … we invite your thinking, but we also announce that there is an eighty per cent chance that we will throw it away and make sure that it is completely wasted.” 20 But there are important differences; in case of design crowdsourcing, it is not just a handful of selected studios competing for one prestigious job that will eventually be paid properly. Instead, hundreds of designers complete a badly paid job beforehand, there is no contract and no client afterwards. But Koolhaas has an important point here: the ethical problem lies not only in the low average wages but especially in the systemic waste of effort and creativity.
There are other crowdsourcing models, also in the design world, in which the contributors become shareholders of the products they help to create (Quirky.com) 21 and others, in which the cash rewards in a contest are significantly higher and the community decides who will get them (Jovoto.com), or where the community works on partly social projects (OpenIDEO.com). In other words, there are possibilities to at least mitigate the hardship of crowdsourcing to some extent. A system such as that of 99designs, however, in which the workers have to gamble for their remuneration, where they have a 1% chance to get paid for their labour while the organisers make a 40% revenue in 100% of the cases, can only be called exploitative and unethical, and the fact that the true price calculation is hidden makes this even more clear. There are initiatives such as No!Spec (no-spec.com) that try to prevent designers from participating in so called speculative work, but it is unlikely that these modes of production are going away. There is just too much profit to be made by the platform owners and too much desperation or naivety among those who participate. Even if Christopher Otey should win his case against CrowdFlower, a national class action lawsuit will not be enough against such a global phenomenon, especially if the crowd chooses to be exploited in that way, instead of revolting against it.
[This is a revised and condensed version of a paper that was originally published as part of the research track of the transmediale 2013 BWPWAP – Back When Pluto was a Planet – and can be found here: A Peer Reviewed Journal About #BWPWAP.]
- Biewald, Lukas. Lukas Biewald of CrowdFlower on TWiST #154 Bonus. YouTube video, 22 June 2011. ↩
- Massolution. Enterprise Crowdsourcing Research Report. Massolution.com, 2012. Page 19. ↩
- Carr, Nicholas. The economics of digital sharecropping. ROUGH TYPE, May 2012. See also Carr: Sharecropping the long tail. ROUGH TYPE Dec. 2006. ↩
- Toffler, Alvin. The Third Wave. New York: Bantam Books, 1989. ↩
- Bruns, Axel. Blogs, Wikipedia, Second life, and Beyond: From Production to Produsage. New York: Peter Lang, 2009. ↩
- Kücklich, Julian. Precarious Playbour: Modders and the Digital Games Industry. The Fibreculture Journal FCJ-025.5 2005. ↩
- Leadbeater, Charles. We-think: Mass Innovation, Not Mass Production. London: Profile Books, 2008. ↩
- Keen, Andrew. The Cult of the Amateur: How Today’s Internet is Killing our Culture. New York: Doubleday/Currency, 2007. ↩
- Grossman, Lev. You — Yes, You — Are TIME’s Person of the Year. Time 25 Dec. 2006. ↩
- Le Bon, Gustave. The Crowd: A Study of the Popular Mind. Minneapolis, MI: Filiquarian Publishing, 2005. Page 10. ↩
- Surowiecki, James. The Wisdom of Crowds: Why the Many are Smarter Than the Few. London: Abacus, 2005. From the introduction. ↩
- Howe, Jeff. The Rise of Crowdsourcing. Wired 14 June 2006. See also: Crowdsourcing: How the Power of the Crowd is Driving the Future of Business. London: Random House Business, 2009. ↩
- Shirky, Clay. Cognitive surplus: How Technology Makes Consumers into Collaborators. New York, Penguin Books, 2011. Page 27. ↩
- McGonigal, Jane. Gaming Can Make a Better World. TED Talk, 2010. ↩
- Herz, J. C. Harnessing the Hive: How Online Games Drive Networked Innovation. Esther Dyson’s Release 1.0 18 Oct. 2002. ↩
- Bogost, Ian. Gamification Is Bullshit Bogost.com Aug 2011. And Persuasive Games: Exploitationware, Gamasutra May 2011. ↩
- Von Ahn, Luis at the MIT Center for Collective Intelligence. See also: Von Ahn, Luis et al. reCAPTCHA: Human-Based Character Recognition via Web Security Measures. Science 321.5895 (2008): 1465–1468.
And von Ahn, Luis. Human Computation. Talk at Library of Congress, YouTube video, 2009. ↩
- Wales, Jimmy. As Wikipedia moves to S.F., founder discusses planned changes. San Francisco Chronicle, Nov. 2007. ↩
- Bernstein, Michael S. et al. Soylent: a Word Processor with a Crowd Inside. Proceedings of the 23nd Annual ACM Symposium on User Interface Software and Technology. New York, NY, USA: ACM, 2010. 313–322. ↩
- Hustwit, Gary. Urbanized. PlexiFilm, 2011. Koolhaas interview at min 51:50. ↩
- Schmidt, Florian Alexander. Hive: From the production for the masses to design by the masses – crowdsourcing und crowdfunding in product design. In bauhaus Vol. 3.: Dinge/Things. Leipzig: Spectormag GbR, 2012. Article published online under: Quirky: Product Design by the Masses. ↩