The Unregulated Work of Mechanical Turk

Nancy Folbre is an economics professor at the University of Massachusetts, Amherst. She recently edited and contributed to “For Love and Money: Care Provision in the United States.“

Ever wonder what our labor market would look like without minimum wages or labor law protections Take a look at the brave new world of online piecework platforms, like Amazon’s Mechanical Turk, which allows employers, politely termed “requesters,” to post jobs for a “global, on-demand, 24 x 7 work force.”

Workers are offered pay for completion of a series of Human Intelligence Tasks (HITs), easily fragmented activities (like transcription, categorization or tagging) in which computers actually need assistance from carbon-based life forms like ourselves.

Spamming and fake reviewing can be easily commissioned. For instance, I could probably pay less than 10 cents apiece for unique posted comments of at least 50 words including at least two positive superlatives on this blog. (Should I discuss this with my editors)

Estimates of what workers can earn on these crowdsourced tasks range from about $1.20 to $5 an hour without any benefits. Employers treat them as independent contractors not covered by federal minimum-wage legislation. A standard terms-of-use agreement gives employers the freedom to reject an employee’s work on any grounds; workers (oops, I mean contractors) have no easy recourse.

Mechanical Turk takes its name from an 18th-century hoax featuring a man-size Turkish puppet that could vanquish most opponents at chess with serene equanimity. Years later it was revealed that his chess table concealed a human prodigy who could manipulate the pieces from underneath with magnets. Other successful companies have adopted equally poetic names, like CrowdFlower and CrowdCloud.

What started as a niche experiment has become a major global industry. Like some other activities, like work at call centers, digital piecework represents a form of virtual labor migration that denationalizes employment. Research by Panos Ipeirotis, a computer expert at the Stern School of Business at New York University, estimates that Mechanical Turk alone engages 500,000 active workers in more than 100 countries, with workers heavily concentrated in two countries: the United States (with 50 percent of the total) and India (with 40 percent).

About 70 percent of its employees are women, many of whom probably can’t find other opportunities to work from home with flexible hours and are therefore willing to accept low wages.

The Mechanical Turk Web site promotes itself with a quotation from a proud chief executive: “Over all, we estimate saving 50 percent over other outsourcing methods.” Yet as both Zakia Uddin on Alternet and Julian Dobson on The Huffington Post point out, these labor practices haven’t gotten as much attention as sweatshop practices in other countries.

A recent Utne Reader article by the California journalist Ellen Cushing briefly profiles some of the industry’s fans as well as its critics. In general, computer scientists, including Professor Ipeirotis, seem quite cheerful about its prospects for improving both efficiency and opportunities for people in developing countries who can gain access to computers.

Low-quality wages may elicit low-quality work. But as Professor Ipeirotis points out, companies can compensate in two different ways, through redundancy (hiring several workers to do the same job and comparing their results) or through use of “gold data” â€" questions to which employers already know the answer, randomly inserted as a test of worker competence.

One recent academic paper on the future of crowd work, acknowledging sweatshop anxieties, asks, “Can we foresee a future crowd workplace in which we would want our children to participate” It does not provide a clear answer.

Such a future can clearly be imagined. But can it be achieved

Workers relying on such low wages and unstable employment are not likely to be able to educate their children enough to escape increasingly high rates of unemployment. A sustainable form of crowdsourcing will require forms of collective governance that mitigate the effects of market competition on those treated as mere links in a chain of algorithmic logic.

In other words, it will require some assurance of human rights, including access to decent employment, living wages and high-quality public education.

Computers don’t care whether they have meaningful opportunities for the development of their potential capabilities. Most humans do. If they had such opportunities, they would not be willing to crawl under a table and create the illusion of an eternally smiling, amazingly intelligent global machine. They would not be willing to get turked.