Policymakers and Agencies Can Better Protect On-Demand Workers From Discrimination

More and more, America’s workers are seeing their jobs deliver less and less of what they need to get by. This is partly because the companies they work for shift risks away from themselves and onto workers, while retaining profits for themselves. In the sector known as the on-demand economy, many online and app-based companies recruit and provide workers who drive, clean, deliver food, do odd jobs, care for children and elders, and perform tasks online—often for very little money, with no job security and no labor or anti-discrimination protections at all. Their employers get away with this major workplace violation largely by classifying their workers as independent contractors (who, by traditional definitions, are not guaranteed the same protections as employees). Unfortunately, these kinds of practices are rapidly expanding: although the on-demand sector is still a small part of the U.S. economy overall, it has grown ten-fold in the last three years.[i]

This guide is intended to assist agency officials and policymakers in ensuring that, no matter how companies choose to label their workforce, workers are protected by anti-discrimination laws.

Are On-Demand Workers Protected From Discrimination On the Job?

Anti-discrimination laws in the workplace exist to protect workers from being treated unfairly and unequally. At the federal level, they prohibit discrimination in the workplace on the basis of race, color, sex, or ethnic origin; age; and disability. State anti-discrimination laws may provide even more protections for workers. Many anti-discrimination laws, such as Title VII of the Civil Rights Act of 1964,[ii] explicitly apply to workers who are classified as employees, but the protections do not generally extend to independent contractors.

The employee versus independent contractor distinction affects many workers at on-demand companies, because rather than classifying their workers as employees, many companies describe their workforce as independent contractors. This is a status that the workers are forced to accept as a condition of employment. In many cases, this label is wrong, and the practice is illegal.[iii] For many, the effect of being classified as an independent contractor means either going without the protections of anti-discrimination laws while on the job, or fighting to be recognized as an employee in order to receive the protections that most workers in the country take for granted.

When on-demand companies treat all workers as employees, workers are able to access the protections of state and federal anti-discrimination laws, as well as other labor protections.

How Might On-Demand Workers Experience Discrimination?

There are several ways that workers for on-demand companies can experience discrimination, frequently without even knowing they have been discriminated against. The scenarios below describe discrimination faced by workers in on-demand companies, who perform duties such as driving, making deliveries, cleaning, and doing odd jobs.

Social science research on the prevalence of gender and racial bias in ratings systems, and anecdotes from online forums and chat rooms where Uber and Lyft drivers swap advice and complaints, point to a problem: on-demand workers are uniquely vulnerable to discrimination, yet lack the vital information that can prove their claim. Lack of transparency and accountability from on-demand companies means workers face overwhelming obstacles to challenge their treatment.

Not Being Hired, or Being Fired

Under Title VII of the Civil Rights Act of 1964, an employer may not make hiring and firing decisions on the basis of race, sex, religion, or national origin.[iv] Just like other workers, workers in the on-demand economy may face discriminatory hiring decisions. However, the hiring and firing—or “deactivation” in the parlance of some companies—process is typically opaque for on-demand companies. According to online driver discussion forums, applicants to Uber or Lyft commonly do not receive an explanation when they have been rejected to drive on the platform.[v]

Uber has acknowledged the existence of rider bias, using such bias as justification for its refusal to allow in-app tipping.

This lack of transparency means that an applicant may be rejected due to his or her race, religion, sex, country of origin, or other grounds that are protected by state and federal anti-discrimination laws, but the applicant would have no way to know the basis for not being hired or for being fired. Such a worker is not only at an information disadvantage (relative to the company) regarding the reason for being rejected, the worker also typically must challenge his or her independent contractor status in order to address the potential discrimination.

In addition to would-be workers potentially being rejected for work on on-demand platforms due to their race, religion, or national origin, customers may also face discrimination. For example, Uber used its notorious “Greyball” system, which presented a potential rider with an inaccurate version of its maps showing where its nearby drivers were, to deny people rides. While much of the recent attention on this program concerned Uber’s attempt to evade law enforcement, it also raises questions about who else was denied service, and why. If a rider were being denied service based on race or religion, the rider would have no way to know that, as the Greyball program operated in secret and would not reveal why a rider was not being picked up by a driver.[vi]

Uber’s regular practices, even outside the Greyball program, also have suggested a pattern of discrimination in customer pick-ups, according to a 2016 study conducted for the National Bureau for Economic Research.[vii] In addition, disability rights organizations have challenged Uber practices that have prevented passengers using wheelchairs or with service animals from accessing the service; meanwhile, Uber and Lyft have claimed that they are technology providers, not transportation providers, and are not subject to the Americans with Disabilities Act.[viii]

Disability rights organizations have challenged Uber’s practices preventing wheelchair-bound passengers from accessing the service.

Bias in the Ratings System

In addition to the potential for discrimination that on-demand workers, like other workers, may face directly from companies in terms of hiring or firing, the very structure of many of these services, whereby workers are rated by customers, can enable biases and lead to discrimination. A 2016 Northeastern University study found evidence of bias along racial and gender lines in two platforms they examined: “On Fiverr, the researchers found evidence that black and Asian workers received lower ratings than white people. And on TaskRabbit, women received fewer reviews than men, and black workers received lower ratings than white ones. Perhaps most troubling, the researchers also found evidence of such bias in the recommendation algorithm on TaskRabbit.”[ix]

There are growing concerns about how customer feedback systems may “hard-wire discrimination into the supervisory techniques of gig economy platforms.”[x] Companies that outsource worker assessments, and ultimately the fates of their workers, to their customers make their workers vulnerable to the enduring prevalence of bias and outright discrimination in society.[xi] For on-demand companies that decide whether to retain workers based on high customer ratings,[xii] such bias by customers may result in on-demand companies firing workers based on race. Title VII does not permit racially motivated decisions by an employer based upon customer preference,[xiii] so on-demand companies should be very concerned about the possibility of this occurring on their platforms.

Not Being Hired Due to an Arrest or Conviction Record

In 2012, the U.S. Equal Employment Opportunity Commission (EEOC) issued guidance on the consideration of criminal records in employment decisions. Because of the “disparate impact” of criminal background checks on people of color, an employer’s use of an individual’s criminal history in making employment decisions may violate the prohibition against employment discrimination under Title VII of the Civil Rights Act of 1964.

Some on-demand employers claim that they are not covered by civil rights and consumer laws that strictly regulate criminal background checks for employment.[xiv] They ignore the fact that many anti-discrimination and consumer protection laws are broadly written to explicitly cover most employment arrangements, including independent contracting.

Hiring policies that place undue emphasis on a criminal record disproportionately impact communities of color, which have been hardest hit by decades of over-criminalization.[xv] Lawyers who represent clients with arrest or conviction records have reported examples of clients being denied work for ride-hailing companies due to arrest records or old convictions, which likely violates Title VII, local fair chance hiring laws, as well as some of the companies’ own stated policies.[xvi]

How Should Agency Officials and Policymakers Address Discrimination in the On-Demand Economy?

Despite the obstacles created by on-demand employers, administrative agencies and legislative bodies can take steps to ensure that on-demand workers and other misclassified workers receive protections from discrimination while on the job.

State and federal administrators and policymakers can focus on enforcing existing laws. In many cases, state and federal laws contain broad definitions that would encompass workers in the on-demand economy as “employees,” no matter whether the company designates them as such. For example, although many on-demand companies call their workers independent contractors, the actual working conditions and arrangements mean that a number of these companies’ workers in fact likely fit within the Title VII definition of an employee.[xvii]

The EEOC announced in its 2017-2021 Strategic Plan that it was adding, as a new priority area, the intention to address “issues related to complex employment relationships and structures in the 21st century workplace, focusing specifically on temporary workers, staffing agencies, independent contractor relationships, and the on-demand economy.”[xviii] The agency stated its plans to clarify the employment relationship and the application of workplace civil rights protections, given the increasing complexity of these employment relationships and structures. At least one claim of race discrimination through the use of app-based rating systems is now pending before the EEOC. Ironically, Uber has acknowledged the existence of rider bias, using such bias as justification for its refusal to allow in-app tipping.[xix]

In addition, some state discrimination laws apply broadly to both employers who discriminate against their employees, as well as workers who are hired under contract. For example, Washington State has an expansive freedom from discrimination statute that protects against employment discrimination and discrimination in public accommodations, and extends its protections to contractors as well as employees.[xx] The Pennsylvania Human Relations Act explicitly protects certain independent contractors from unlawful discriminatory practices.[xxi] State public accommodation laws, and the anti-discrimination provisions within them, may also be applicable to on-demand companies that are providing a service to passengers and drivers.[xxii]

State and local agencies should investigate companies to ensure that they are complying with state anti-discrimination laws.[xxiii] In addition to enforcing laws that already exist to protect workers, state legislatures should resist attempts by on-demand companies to create special exemptions for their sectors from state laws. Lawmakers can also clarify the definitions in their laws to ensure that these companies cannot game the system.[xxiv]

What Can On-Demand Companies Do?

On-demand companies do not need to wait for policymakers to act before they address the discrimination some of their workers experience. For one thing, they could decide to treat all their workers as employees, protected by the same labor standards and anti-discrimination laws as the companies’ engineers and executives.

Even if all workers were employees, that change in status would not fully address the discrimination many on-demand workers experience, so companies need to go further and assess how their customer rating systems can be modified to root out racial, religious, and sexual bias.[xxv] In addition, companies can take steps to address the issues raised in the 2016 National Bureau for Economic Research study, which showed driver discrimination against passengers in the process of accepting ride requests.[xxvi]

Researchers found evidence that black and Asian workers received lower ratings than white workers.

Various researchers have offered suggestions for what companies can do to mitigate the occurrence of bias in rating systems. For example, Alex Rosenblat suggests that Uber and similar on-demand companies that use customer ratings conduct internal audits to assess whether members of protected classes receive systematically lower ratings, and if so, to proactively adjust ratings to make up for the bias the company identifies. Such a recommendation that companies look at their own data reflects the reality that outside researchers would likely be unable to conduct a rigorous independent audit due to lack of access to relevant data. Rosenblat also recommends that, for a given driver with low ratings, companies like Uber should provide a more diverse set of passenger-reviewers, and allow the system to “learn” the rating biases of certain demographics of reviewers and give them an appropriate weight.[xxvii] On-demand companies such as Uber possess enormous technological and creative abilities. There is no reason to assume that they cannot apply their considerable ingenuity to studying and offering real solutions to the vexing and persistent problem of discrimination.

There is no doubt that technology has tremendous potential to improve businesses, consumers, workers, and our economy overall. But the use of new technology should not be an excuse for companies to engage in old-style gaming of the core labor and employment laws that created the American middle class, and that are key to ensuring that working families have the most basic levels of economic security. Instead, as new technology develops, it must be harnessed to build a more inclusive economy—one that delivers to all of America’s workers a secure income, social and legal protections, and the right to engage in collective action.



[i] Diana Farrell & Fiona Greig, Paychecks, Paydays and the Online Platform Economy, JP Morgan Chase & Co. Inst. (Feb. 2016), https://www.jpmorganchase.com/corporate/institute/document/jpmc-institute-volatility-2-report.pdf.

[ii] See 42 U.S.C.A. § 2000e et seq.

[iii] Catherine Ruckelshaus, Nat’l Emp’l Law Project, Independent Contractor vs. Employee: Why Misclassification Matters and What We Can Do to Stop It, (2016), http://www.nelp.org/publication/independent-contractor-vs-employee/.

[iv] 42 U.S.C.A. § 2000e et seq.

[v] See, e.g., http://www.ridesharingdriver.com/rejected-from-uber-and-lyft-can-you-reapply/ (Summarizing typical lack of information from company when a driver is rejected),  https://uberpeople.net/threads/rejected-account.1554/ (Driver struggling to understand why he/she was no longer able to access account), https://uberpeople.net/threads/rejected-by-lyft-baffled.91103/ (Even after pressing for an explanation, prospective Lyft driver was told that they Lyft was “unable to provide you the exact reason” why he was rejected),

https://uberpeople.net/threads/lyft-rejected-my-application-until.80010/ (Lyft support told driver that they cannot tell her why he/she was denied).

[vi] See Adrienne LaFrance, Uber’s Secret Program Raises Questions About Discrimination, The Atlantic, March 3, 2017, https://www.theatlantic.com/technology/archive/2017/03/uber-ghost-app/518610/?utm_source=atltw. The U.S. Department of Justice has begun a criminal probe into this program. See Dan Levine & Joseph Menn, Exclusive: Uber faces criminal probe over software used to evade authorities, Reuters, May 4, 2017, http://www.reuters.com/article/us-uber-tech-crime-exclusive-idUSKBN1802U1.

[vii] See Faiz Siddiqui, Your Uber driver is twice as likely to cancel if you’re black, The Washington Post, Nov. 1, 2016, https://www.washingtonpost.com/news/tripping/wp/2016/11/01/your-uber-driver-is-twice-as-likely-to-cancel-if-youre-black/?utm_term=.24c33e91f0b9; See also Jennifer Stark and Nicholas Diakopoulos, Uber seems to offer better service in areas with more white people. That raises some tough questions, The Washington Post, Mar. 10, 2016, https://www.washingtonpost.com/news/wonk/wp/2016/03/10/uber-seems-to-offer-better-service-in-areas-with-more-white-people-that-raises-some-tough-questions/?utm_term=.4a9ed52462e7.

[viii] See Michael Tarm, Disability rights group sues Uber over wheelchair access,  Associated Press, October 13, 2016, https://www.usnews.com/news/business/articles/2016-10-13/disability-rights-groups-sues-uber-over-wheelchair-access; Lori Aratani  “Blind woman sues Uber, says she and her guide dog were refused service, The Washington Post Jun. 3, 2016, https://www.washingtonpost.com/local/trafficandcommuting/blind-woman-sues-uber-says-she-and-her-guide-dog-were-refused-service/2016/06/02/7391156e-28f8-11e6-ae4a-3cdd5fe74204_story.html?utm_term=.df5413b2104a.

[ix]  Will Knight , Is the Gig Economy Rigged?, MIT Technology Review, Nov. 17, 2016,


[x] Noah Zatz, Beyond Misclassification: Gig Economy Discrimination Outside Employment Law, On Labor, Jan. 19, 2016, https://onlabor.org/2016/01/19/beyond-misclassification-gig-economy-discrimination-outside-employment-law/.

[xi] See Alex Rosenblat et al., Discriminating Tastes: Customer Ratings as Vehicles for Bias, Oct. 19, 2016. Available at SSRN: https://ssrn.com/abstract=2858946.

[xii] See Samantha Allen, The Mysterious Way Uber Bans Drivers, The Daily Beast, Jan. 27, 2015, http://www.thedailybeast.com/articles/2015/01/27/the-mysterious-way-uber-bans-drivers.

[xiii] Chaney v. Plainfield Healthcare Ctr., 612 F.3d 908 (7th Cir. 2010).

[xiv] E.g., Gillette v. Uber Technologies, information available at http://gbdhlegal.com/cases/mohamed-et-al-v-uber-technologies-inc/; the consolidated complaint has been designated In re Uber FCRA Litigation, No.  3:14-CV-05200. Uber and its background check company have reached an agreement with job applicants to dismiss allegations the companies violated federal and state background check laws, and Uber agreed to pay $7.5 million to the settlement class members. Rachel Graf, “Uber, Background Check Co. Settle Job Screening Suit,” Law360, May 30, 2017, https://www.law360.com/articles/929478.

[xv] See section on disparate impact discrimination and criminal records, and arrest and conviction disparities by race, EEOC Enforcement Guidance, Consideration of Arrest and Conviction Records in Employment Decisions Under Title VII of the Civil Rights Act of 1964, (April 25, 2012).

[xvi] E.g., Emily Hoffman, Staff Attorney with Community Service Society of New York described in email correspondence on March 27 and June 6, 2017 the experiences of two clients who were denied the opportunity to drive for UberEats based on past convictions, without the company having performed the required analysis under New York City’s Fair Chance Act. One of Ms. Hoffman’s clients had a conviction from 1978, and received his license from the city agency that regulates for-hire drivers in New York City, but was still denied the opportunity to drive for UberEats based on his decades-old conviction. See also examples in “Ensuring Fairness in Background Checks,” June 2016,  http://www.nelp.org/content/uploads/Policy-Brief-Fairness-Background-Checks-On-Demand-Work.pdf, page 4; see also “Outten & Golden LLP: Lyft, Inc. Accused of Violating Philadelphia “Ban the Box” Law,” March 15, 2017, http://www.outtengolden.com/outten-and-golden-lyft-inc-accused-of-violating-philadelphia-ban-the-box-law.

[xvii] In most instances, an individual performing labor or services for another should be covered as an employee under our employment laws, unless the person operates an independent business, with specialized skill, capital investment, and the ability to engage in arms-length negotiations over the terms of a job. See Independent Contractor vs. Employee, supra fn. 3. In addition, recent decisions by state agencies that use similar “right to control” standards for determining who is an employee as Title VII case law have concluded that the facts of the relationship between drivers and companies like Uber lead to an determination of employee status. For example, New York’s Department of Labor  determined that two former Uber drivers were employees eligible for unemployment payments, Noam Scheiber, Uber Drivers Ruled Eligible for Jobless Payments in New York State, New York Times, Oct. 12, 2016, https://www.nytimes.com/2016/10/13/business/state-rules-2-former-uber-drivers-eligible-for-jobless-payments.html?_r=0.

[xviii] U.S. Equal Employment Opportunity Commission Strategic Enforcement Plan
Fiscal Years 2017 – 2021,  https://www.eeoc.gov/eeoc/plan/sep-2017.cfm.

[xix] Dan Adams, Boston-based attorney argues Uber’s star ratings are racially biased, Boston Globe, Oct. 7, 2016, https://www.bostonglobe.com/business/2016/10/06/attorney-for-uber-drivers-says-star-ratings-are-racially-biased/R28mqWL6ShjMFB5xAr3uGL/story.html.

[xx] Wash. Rev. Code Ann. § 49.60.030 (West); Currier v Northland Servs., Inc., 182 Wash. App. 733, 744 (2014).

[xxi] 43 Pa. Stat. Ann. § 955 (West); Velocity Express v. Pennsylvania Human Relations Comm’n, 853 A.2d 1182, 1186 (Pa. Commw. Ct. 2004).

[xxii] E.g., Cal. Civ. Code § 51 (West), Alch v Superior Court, 122 Cal. App. 4th 339, 391 (2004); Conn. Gen. Stat. Ann. § 46a-63, § 46a-64 (West); N.J. Stat. Ann. § 10:5-1 et seq. (West), Nat’l Org. for Women, Essex Cty. Chapter v. Little League Baseball, Inc., 127 N.J. Super. 522, 531 (App. Div.), aff’d sub nom. Nat’l Org. of Women v. Little League Baseball, Inc., 67 N.J. 320, 338 A.2d 198 (1974).

[xxiii] See Independent Contractor vs. Employee, supra fn. 3, for other strategies that state and local agencies can employ, including interagency task forces, procurement rules, and community collaborations.

[xxiv] See Nat’l Emp’l Law Project, The On-Demand Economy and State Labor Protections, January 11, 2017, http://www.nelp.org/publication/the-on-demand-economy-state-labor-protections/.

[xxv] For additional suggestions for on-demand companies, see Ellen Powell, How can the gig economy address discrimination?, The Christian Science Monitor, Nov. 2, 2016, http://www.csmonitor.com/Business/2016/1102/How-can-the-gig-economy-address-discrimination and Luke Stark, Recognizing the Role of Emotional Labor in the On-Demand Economy, Harvard Business Review, Aug. 29, 2016, https://hbr.org/2016/08/recognizing-the-role-of-emotional-labor-in-the-on-demand-economy.

[xxvi] See Nov. 2, 2016 letter from Sen Franken to Uber and Lyft, asking what specific steps the companies were taking to address discrimination against riders. https://www.franken.senate.gov/files/letter/161102_UberLyft.pdf.

[xxvii] Alex Rosenblat, Uber’s Pax: Hidden Bias in Rating Systems, Dec. 30, 2015, Working Paper prepared for CSCW Workshop: Algorithms at Work https://algorithmsatwork.files.wordpress.com/2016/02/rosenblat-uber_s-pax-bias-in-rating-systems-cscw-2016.pdf.

Back to Top of Page