Crowdsourcing: Do the benefits outweigh the low-wage pay?

WEST LAFAYETTE, Ind. – A millennial wakes up late for work. Usually she takes the bus, but she has an early meeting so she requests an Uber ride. After her meeting, she realizes she forgot to pick up a dessert for a dinner party tonight. It's a real bad case of the Mondays. She uses Task Rabbit to request someone to pick up a cake from a local baker and deliver it to her apartment by 5 p.m. Later in the day, she browses Airbnb to find a place for her upcoming vacation to Cancun.

She can do all of this because of crowdsourcing, a process of connecting a "crowd" of workers to those in need of a service. Its convenience and ease of use is driving huge growth for companies that know how to harness the power of crowdsourcing. 

But crowdsourcing also is being criticized for its negative effects on workers. For instance, Uber. The company gained popularity quickly because it offered an alternative to taxis, which had little reason to innovate without competition. However, to stay competitive Uber slashed rates for customers, which dropped wages for drivers.

The economy of crowdsourcing is essentially unregulated. Uber drivers are considered independent contractors and, therefore, must incur all of the expenses that go along with a taxi service, such as car maintenance. Nor is Uber obliged to follow minimum wage laws.

Purdue professor Alex Quinn is hopeful about the future of crowdsourcing and the so-called "sharing economy," but he's also wary of what unfettered innovation might bring. He will explore the future of crowdsourcing and its pitfalls during his presentation at Dawn or Doom '16, a conference on the risks and rewards of new technology at Purdue University. Dawn or Doom will be held Oct. 3-4, on the Purdue West Lafayette campus and is free and open to the public.

Technology brings the workers and those in need of the work together, but human factors still play a large role in the success or failure of an online interaction, as well as in how it benefits, or doesn't, both parties, says Quinn, an assistant professor in Purdue’s School of Electrical and Computer Engineering.

For one thing, as humans we tend not to treat each other with the same regard remotely as we would in person. Even theoretically enlightened university researchers started crowd-sourcing with extremely low rewards for respondents.

"You would be surprised at how much some people want you to do for 8 or 10 cents," Quinn says. "There's certainly a tendency, formerly of researchers, to just really shoot for the floor."

He says standards should be set and enforced from the very beginning when the technology is being researched and created.

"We have to design our technology and experiments a little differently, knowing that human effort really is precious," Quinn says.

Quinn and his team of students also are working on ways to improve the crowdsourcing process. They have three projects, each dealing with a different problem that crowdsourcing presents: instruction quality, duplication of results and division of labor.

One of the most difficult problems for crowdsourcing is how to split up tasks in such a way that workers can come and go from a project at its various stages without disrupting progress. Quinn and his students developed a method of chopping up a question into a "space," so all of the parts come together to equal a comprehensive set of answers.

Take the question: What are all of the ways you can use a brick? If only one person answers, it will be an incomplete set of data. But if 100 people help answer, a more comprehensive dataset can be produced. First, you pose the question and gather the responses. The next step is to categorize the answers, for example nonviolent or violent uses of bricks. Then, the answers are arranged in a logical way.

To apply that to the real world, a political candidate could crowdsource every possible debate question or an economist who needs to classify thousands of reports into a set number of categories could employ the crowd to help.

Quinn knows that using crowdsourcing can save precious time. When he was finishing his own dissertation, he tapped the crowd to format over 200 references, so he could focus elsewhere.

"Everyone should be doing that," Quinn says. "When you are in the final throes of submitting your dissertation or whatever it is, you should be able to get help freely. If there are people out there willing to help, they ought to be able to help." 

Writer: Kirsten Gibson, technology writer, Information Technology at Purdue (ITaP), 765-494-8190, gibson33@purdue.edu 

Source: Alex Quinn, aq@purdue.edu

Note to Journalists: Assistant rofessor Alex Quinn will be speaking at Dawn or Doom’16 at 2:15 p.m. Oct. 4 in Purdue's Stewart Center. Press registration is available at http://www.purdue.edu/dawnordoom/Press