Research shows that algorithms using sociodemographic data (e.g., race, gender, education, etc.) can produce biased outcomes that cause many consumers to be excluded from or endure lower levels of service. Though research suggests that these algorithms are more profitable than unbiased algorithms that do not use sociodemographic data, prior findings do not consider potential social effects of these algorithms on consumer demand.

This research investigates the dynamic outcomes of competition between biased and unbiased algorithms in a market where word-of-mouth influences consumer choice behavior. Relative to unbiased algorithms, this research demonstrates that biased algorithms can be more profitable in the short run but less profitable in the long run, due to consumer word-of-mouth.

Models and simulations show that word-of-mouth leads marginalized consumers to gravitate towards easier-to-access unbiased algorithmic services. Non-marginalized consumers, on the other hand, learn they have a relatively easier time accessing services anywhere.

When sufficient numbers of marginalized and non-marginalized consumers learn from each other via word-of-mouth, long run demand is greater for unbiased algorithmic services. This research demonstrates that firms that use unbiased algorithms and account for social effects (e.g., word-of-mouth) in the algorithm’s design can reduce algorithmic bias while improving both long-term profits and societal well-being.

Because of growing concerns about responsible use of artificial intelligence, European and US regulators recently introduced legislation to protect consumers from algorithmic bias. These policies hold firms accountable for the fairness of their algorithms while relying on consumers to report when unfairness occurs. Our research reveals unintended consequences of these policies resulting from differences between how firms and consumers assess fairness.

Current algorithmic fairness standards use firm data and measures of statistical parity to determine if demographic groups are being treated similarly. However, the average consumer does not have access to firm data. Consumers are left to assess fairness by gathering information from their social networks about the firm's service treatment. We model how consumer assessments can trigger the spread of unfairness beliefs, even when the firm's algorithm is fair.

We show that a lack of algorithmic accountability may lead consumers to paradoxically believe that a firm with an unfair algorithm is less biased than a firm with a fair algorithm. We also demonstrate how accountability via a third-party watchdog institution could reconcile these different perceptions of fairness and provide both firms and consumers with better insights into algorithmic bias.

We investigate how competition for reputation among consumers can impact the effectiveness of firm interventions. In situations where the firm attempts to intervene to change consumer behavior (e.g., advertising, policy guidelines, legal threats), the consumer's response may depend not only on the consumer's own reputation-building considerations but also on those of others.

In this study, we model uploaders' decisions to enhance their reputations by uploading pirated content to a digital platform. Uploaders weigh the reputational benefits from uploading versus the costs from penalties associated with copyright lawsuits (firm interventions). Furthermore, competition from other uploaders who also seek to build their reputations impact each uploader's potential gains.

Using a novel data set, we find empirical support for our conceptual framework, which suggests copyright lawsuits may deter uploading in the short-run but may, in some cases, lead to more piracy over the long-run. Our conceptual framework proposes that low-reputation uploaders decrease their reputation-building activity in response to lawsuits.

However, this decrease may create an opportunity for their high-reputation competition to increase their uploading activity and enhance their reputations. The implication for firms is that when considering consumer interventions to mitigate harmful activity, firms need to account for consumer reputation concerns and plan accordingly.

Parents want the best possible education for their children and the ability to choose their schools. However, prior research shows that as the ability to choose schools increases, so does school racial segregation. Segregation reduces educational, economic, and health outcomes for all students. Because of these negative outcomes, US policymakers seek to increase integration, often inadvertently reducing a parent's ability to choose.

This research proposes that education marketing interventions can increase integration rates while maintaining a parent's ability to choose. We propose that information that counters parents' prior beliefs about a racially-different school (i.e., a school whose majority race differs from the child's) can increase a parent's likelihood of choosing the school. This research uses machine learning methods and a series of experiments to identify parents' baseline beliefs and motivations for selecting or avoiding racially-different schools.

We then test marketing interventions for efficacy in influencing over 2,400 parents' choice of a racially-different school. We find that providing information that counters prior beliefs is up to 33 times more effective in motivating a White parent's choice of a primarily-Black school than a Black parent's choice of a primarily-White school. School administrators and policymakers may benefit from using these marketing approaches.

With increased competition in the attention economy, content creators need sustainable strategies for cultivating consumer engagement. One oft-advocated strategy is to develop relationships with audiences. In this work, we explore how the use of first-person singular pronouns (FPP), such as ``I'' or ``me,'' can create the feelings of personal closeness that build relationships and generate engagement.

Combining experimental and field data from multiple domains, including social media, news and message boards, we show that FPP increases relationship-motivated engagement (including votes, likes, and comments) in initial interactions. However, because relationships are dynamic, we expect the effects of FPP on engagement to wane over time, as their continued use suggests self-focus, rather than relationship motivations, to audiences.

Using data from multi-year studies of content creators on Reddit and Twitter, we show that the positive effects of FPP attenuate with time, and can even have negative effects on engagement. Our results demonstrate the potential of short-term as well as long-term pronoun usage in content engagement strategies, and suggest that content creators must consider the stage of their relationships with their audience members and their mix when using FPP to generate engagement.

News about biased algorithms has made people realize that decisions made by artificial intelligence systems can be influenced by demographic details like race, ethnicity, and gender. Often, people hear about these decisions—like getting a loan, insurance, or school admission—through friends or family who share similar backgrounds.

For instance, if women hear that other women are being denied credit cards by a company, they might assume that gender plays a role in these decisions and decide not to apply to that company, even if they don’t know why the others were denied. This study explores how these shared stories about acceptance or rejection by services can shape people’s choices, especially when it comes to services that screen applicants.

We find that this specific type of word-of-mouth, which we call service acceptance word-of-mouth, can guide historically marginalized consumers towards companies that use fair algorithms and non-marginalized consumers towards companies using unfair algorithms. Interestingly, we also find that when people from different backgrounds share and value each other’s experiences, they might all choose the same firm—even if the firm's algorithm is not in everyone’s best interest.

We demonstrate that this shared choice can arise from people trying to maximize their own benefits, even without any explicit concern for social justice, fairness, or morality.