DATING

According to many prominent theories of justice, social institutions that exercise significant power over individuals must be justified. They can be so justified if they exercise their power in line with individuals’ interests (Rawls Reference Rawls1971; Raz Reference Raz1986; Scanlon Reference Scanlon1998). Many important debates in recent years have concerned which institutions this idea applies to. Is the family a social institution (Schouten Reference Schouten2019)? Is the institution of marriage (Macedo Reference Macedo2015)? Are private corporations (Cordelli Reference Cordelli2022)? The issue of how we should view the seemingly more informal practices by which people initiate intimate relationships has not been explored. In different historical periods and cultural settings, the degree to which practices for intimacy initiation have been formalized has varied greatly. In the US, “dating” refers to a more formalized set of activities, in the sense that this practice involves specific chains of events, while in European countries, the process for intimacy initiation tends to be less rigid.Footnote ⁵ For reasons of simplicity, I will refer to all forms of initiating intimate relationships and interactions (including so called hookups) as “dating” unless specified. This includes but is not limited to the specific American practice of going on dates, the “dating game,” which developed during the 20^th century (Weigel Reference Weigel2016).

Since dating apps are part of the wider practice of dating, we need to explore whether dating is a social institution exercising power over people in the relevant way, before we can ask how dating app power could be exercised in line with individuals’ interests. In other words, we need to establish that dating is the kind of phenomenon to which the demands of justice apply.

While Rawls, for instance, theorizes about how social institutions should govern and be governed, he does not have a full account of what a “social institution” is (Miller Reference Miller2019). In social science and philosophy of social science, however, a social institution is often defined as a social system or structure organizing the primary social practices, roles, and relationships within a culture. For instance, Turner (Reference Turner1997, 6) argues it is:

We can use the term “social institution” to refer to institutions at varying levels of abstraction. Thus, democracy is a social institution, but so is the Swedish parliament. Dating is not like democracy, in the sense that the latter has a formal structure with a defined role distribution and formalized rules about authority within the institution, while the former often lacks such a formal process. Having said this, it is important to note that the evaluation of dating is made more difficult by the fact that there is a mystification embedded in the practices of the intimate spheres (Illouz Reference Illouz2012, 8–9)—we are primed by convention to think about intimate relationships as the products of chance or fate (Bergström Reference Bergström2013, 6; Slegers Reference Slegers2021). According to predominant narratives, dating is a series of events that just so happen to lead us to end up as a couple, which just so happens to be the kind of union that almost everyone else in our society is either a member of or trying to enter (Bergström Reference Bergström2022, 71–4). We are primed, especially, not to think about dating as a rule-bound process, influenced by the economic, social, or cultural structures of the rest of society (Bergström Reference Bergström2022, 8). As a result, dating is to some extent an invisible institution. Yet, societies tend to have some established practice for the initiation of intimate relationships. And these practices tend to be vital to the very function and reproduction of society over time. Indeed, they are examples of solutions to the aforementioned fundamental problems with “reproducing individuals,” “sustaining viable societal structures,” and “producing life-sustaining resources”—in particular that of human care and care work.

To see this, consider that in the study of social institutions, some institutions are considered particularly fundamental, meaning that they are presupposed by other major institutions and necessary for society’s functioning over time. The family is often understood to be a prime example (Lamanna Reference Lamanna2002; see also Okin Reference Okin1989). In our societies, all other institutions rely on the care work and social reproduction that happens in families (Kugelberg and Kugelberg Reference Kugelberg and Kugelberg2024). This is something recognized by social scientists and economists (see, for instance, Becker Reference Becker1981), and a broad class of normative theorists including feminists (Fraser Reference Fraser2014), Marxists (Federici Reference Federici1975), and liberals (Okin Reference Okin1989; Rawls Reference Rawls2001). Citizens, consumers, and workers: they generally enter the world through the family. Capitalism, the university, the parliament: they all presuppose the work of the family.

This fundamental institution in turn relies on intimacy initiation, in some form. In some historical periods, this formation of care work units has been referred to as courtship. In other periods, arranged marriage has been the norm. Where, as in the US, the American form of dating is an established practice for intimacy initiation, it is a precondition for the formation of socially prescribed forms of care work units such as couples and families (Weigel Reference Weigel2016). This does not imply that this form of dating, or this kind of care work unit, is necessary for any society as such—there are many alternative ways in which intimate relationships could be initiated. And there are many possible ways in which societies are and could be reproduced.

However, any society needs some way for individuals to form the relationships in which necessary care work and social reproduction can take place. In a society where dating serves this function, it is a fundamental social institution. Even though dating does not have the same formal role distribution as the Swedish parliament, it does have a clear social function, as well as an informal yet clearly socially enforced role distribution. Like schools and parliaments, dating also contributes to the reproduction of other important institutions. This does not entail that individuals engaging in this institution do so with the intention of contributing to this reproduction. The same is true for other institutions; to take an example, many shop at Amazon.com without having as their intention to reproduce global capitalism (Agmon Reference Agmon2022, 33–4).

The rituals, role distributions, and social norms of dating are under constant negotiation. In the US, it is, for instance, no longer the case that only members of the opposite sex can date each other, or—at least sometimes—that men must pay for everything consumed. However, the fact that social norms have changed is not an argument against seeing dating as a social institution because social institutions generally do not remain stable. Consider, for example, how family arrangements and our parliaments have changed under the pressure of social and technological change, as well as moral and political arguments.

Note that I am not making a normative argument about the point of dating, nor do I claim it is or should be about contributing to society by having children, caring for workers, or building a home. Many have questioned these norms, and rightfully so. Importantly, the norms regulating dating are questioned in part because people see its influence over them—because they take it seriously and see the immense power this institution has over their lives.

So much for dating. Where does this leave dating applications?

DATING APPS

According to sociologist Eva Illouz (Reference Illouz2012, 180), “online dating represents the most significant trend in modern courtship.” To understand the relationship between dating apps and dating, we need to consider how the development of online dating took place during a period in history that saw wider social changes. One of these is an ongoing compartmentalization trend whereby various social activities and spheres are each given their distinct logic: the gym is for moving one’s body and colleagues are for work. While before, the initiation of intimate relationships happened where people worked and lived, they now meet in spaces specifically designed for initiating intimate relationships. Instead of farmers meeting during harvest, fishermen falling in love by the sea, and students getting together at school, sex and intimacy have been assigned a specific time and place: the dating app (Bergström Reference Bergström2022, 171–3; Illouz Reference Illouz2012, 180–3; Kaplan and Illouz Reference Kaplan and Illouz2022, 26–7; Williams Reference Williams2024, ch1).

Through this greater separation between sex and the rest of the social world, we are approaching an increasingly “pure” sexual sphere. This means, first, that dating qua social institution becomes visible. It is now easier to see that dating is not a random series of quasi-magical events—that love is not (only) the work of Cupid’s arrow. It is now easier to recognize that dating is a social institution, with its own governing norms. It becomes clearer that many of the problems that pester other social institutions are pestering institutions for intimacy initiation as well. Since dating apps have generated millions of experiences and vast amounts of data, it is now harder to dismiss, ignore, or romanticize problems that were perhaps always there (Bedi Reference Bedi2019; Bergström Reference Bergström2022; Slegers Reference Slegers2021; Weigel Reference Weigel2016).

Second, this development underscores the point that the sexual sphere needs special attention. Given that in contemporary practice, the sexual sphere increasingly is separated from other areas of human life, there is a need for the construction of particular principles of social justice for the activities that happen within it (compare, James Reference James2012, 105; Reference James2014, 114). As Rawls (Reference Rawls1971, 29) argues, “the correct regulative principle for anything depends on the nature of that thing.”

Regulative principles may apply to different agents in different ways. The state ought to bring about conditions in which people’s claims can be realized. Individuals have duties of justice: they should contribute to the reform of unjust norms and institutions and the maintenance of just ones. Where just institutions are lacking, they should contribute to bringing them about (Shelby Reference Shelby2007). I say more about this in the next section. But what about the duties of private intermediaries, such as dating apps? To answer this question, we must first properly capture the role dating apps play in dating as a social institution.

Illouz (Reference Illouz2012, 170–84) argues that while dating apps promise to rationalize the sphere of romance, these “technologies of choice” are not neutral: in importing the tools and logic of the market (Agmon Reference Agmon2024), they affect the shape and function of both intimacy initiation and intimacy itself.Footnote ⁶ Dating apps, thus, can be seen as both a response to, and drivers of, the compartmentalization of the sexual sphere (Bergström Reference Bergström2022, 39–57). Against this background, they are best understood as agents intervening in the social institution of dating.Footnote ⁷ As intermediaries, dating app corporations profit from mediating the relations, filling the roles, and distributing the social goods of the sexual sphere that are organized through the institution of dating, an institution which is a precondition for other fundamental institutions of our societies. Inserting themselves as middlemen in dating enables dating apps to shape what is possible within individuals’ intimate relationships.Footnote ⁸ In the context of social media apps in the digital public sphere, Seth Lazar (Reference LazarForthcoming) argues that intermediary power:

Dating app companies, similarly, govern the digital sexual sphere in several ways. Acting as middlemen between individuals who otherwise would not have come into contact, these companies cannot avoid exercising some power over users. They must design their platforms and sort their information in some way, according to some guidelines and rules.Footnote ⁹ Not influencing users’ intimate lives does not exist as an option. Dating apps will influence what happens; the question is how this influence can best protect individuals’ interests—rather than just the interests of shareholders.

In sum, a handful of private corporations have grabbed substantial control over dating. They oversee how this social institution, and its norms, develop. By comparison, the rest of us have very little say in what practices for intimacy initiation should look like. There is reason to question this situation. When social institutions have great power over our lives, we can at the very least demand that they work well for everyone, and not only for those already well-off. This is what I will argue concerning dating. We should demand that dating apps exercise their power in a way that does not undermine our interests.

I will now say more about the claims individuals have in the sexual sphere.

THREE CLAIMS IN THE SEXUAL SPHERE

Outside of the sexual sphere, liberal egalitarians often argue we have three broad claims on others: (1) that others do not wrongfully interfere with us (noninterference), (2) that we stand in relationships of equality with them (equal standing), and (3) that they act to improve our circumstances of choice (choice improvement) (compare, Kolodny Reference Kolodny2023, ch1).Footnote ¹⁰ These claims have not previously been extended to the context of sexual attention or the initiation of intimate relationships in the sexual sphere. In what follows, I will assume that versions of these claims do apply in the sexual sphere. While the present context does not let me elaborate on this assumption fully, I will point to a number of features of the sexual sphere that suggest these claims—or something like them—are applicable in relation to the initiation of intimate relationships.

ARCHITECTURE

In the context of platform technology, “architecture” refers to the design of the platform and its protocols (Van Dijck, Poell, and De Waal Reference Van Dijck, Poell and De Waal2018). Like the architecture of a physical space, the app’s architecture affects where users can go and what they can do in the digital space. In the context of dating, the architecture does not fully determine with whom users will match or go on a date, but it nudges their choices by providing the framework through which specific option sets are presented and other options made impossible.

One major dating app architecture is the match architecture, in which users are presented with potential partners, whom they then can choose whether to accept (David and Cambre Reference David and Cambre2016). Importantly, they can only start writing to someone who, at that point, has already indicated an interest in them (Wu and Trottier Reference Wu and Trottier2022, 91). This furthers users’ interests in noninterference, as they do not risk being contacted without their consent. The grid architecture allows people to send a direct private message to anyone whose photo appears on the screen. Both architectural designs solve the Information Problem, to varying degrees. On the match architecture, there is no ambiguity regarding the intent of the particular user: they have matched with you. Even if users of the grid architecture cannot know if they have sexual standing in the eyes of the person they start chatting with, they can infer other users are looking for some kind of intimate connection.

Dating apps, importantly, are built around private messaging. Users cannot see who others interact with, or who their friends and family are. Compare this to the digital public sphere, where platforms are built around interactions with public or semipublic posts, as well as the building of networks. Where the issue for actors operating in the digital public sphere is to determine which content gets to go “viral” or be seen by more people, it is a crucial feature of dating apps that they lack publicity. This aspect of dating apps has contributed to the compartmentalization of society and privatization of dating where individuals’ search for potential partners is fully independent of their communities (Bergström Reference Bergström2022, 173–5).

Therefore, the privacy feature of dating app architecture contributes to solving the Social Control Problem by allowing people to follow any sexual life plans even in the presence of harshly enforced norms of chastity, monogamy, or heterosexuality. Shielding them from outside interference from their social network, the app provides people living in the presence of oppressive norms with an opportunity set that is valuable given these norms. Because individuals can chat with a greater number of potential partners without either their social network or these potential partners finding out, it also becomes less costly to take advantage of these opportunities (Bergström Reference Bergström2022, 87–91). Users can develop sexual standing without risking their social standing.

Sexual minorities have historically often not had their sexual-sphere-based claims realized—in fact, the state and important social institutions such as the church have actively tried to undermine them (Rubin Reference Rubin and Vance1984). Therefore, their claims should be taken particularly seriously in the deliberation about potential future regulation as their realization counts as a particularly strong argument in favor of their power being at least partly justified. There are signs that these claims are being better realized. Pew Research Center (2023, 8, 18) reports that “51% of lesbian, gay or bisexual (LGB) Americans say they have ever used a dating site or app, compared with 28% of those who are straight,” and that “LGB users are more likely than straight users to [report positive experiences] (61% versus 53%).” Consider that 24% of partnered LGB adults say they met their partner online dating, compared to 9 per cent of straight Americans (Pew Research Center 2023, 7).

What dating apps provide is not only truly valuable but also something individuals are owed as a matter of justice. They allow individuals to access sexual standing and opportunities to follow sexual life plans without having to worry about social sanctions. However, as I have argued elsewhere (Kugelberg Reference Kugelberg2021, 373; see also Okin Reference Okin1989, 183; Jewkes and Morrell Reference Jewkes and Morrell2010), a feminist approach to targeting gender norms needs to be both “gender-transformative,” as in aiming to alter oppressive norms, and “gender-sensitive” as in providing “those who presently live under such norms with opportunities that are valuable given said norms, and which protects them from harm in the meantime.” The dating apps’ privacy architecture amounts to a gender-sensitive approach because this feature allows women to discreetly navigate the sexual sphere in the presence of harmful modesty norms. However, because users’ norm infringement happens out of sight, this behavioral change will not necessarily be reflected in a change in the social norm itself (Bergström Reference Bergström2022, 163). To change the norm, behavioral changes would need to be public (Bicchieri Reference Bicchieri2017, 1–49; Brennan et al. Reference Brennan, Eriksson, Goodin and Southwood2013, ch1). Doing little to change the idea that modesty is important for women, app architectures generally are not gender-transformative. In the digital sexual sphere, courtship norms remain in place.

Some apps have filters that let users fill in their preferences for potential partners of a certain race or age, for example. This architectural feature offers to help users solve the matching problem by helping users find relevant potential partners while not having to see “irrelevant” ones. This way, users could potentially develop sexual standing faster. However, asking users to predefine their preferences via filters also promotes existing sexual and social hierarchies and encourages people to conform to discriminatory and oppressive sexual norms. Such architectural features run the risk of reinforcing tendencies of homogamy concerning age, looks, race and level of education. When probed, people are likely to define their preferences in relation to people around them, leaving less room for surprise or exploration.Footnote ¹⁷ This is especially troubling in the context of historical injustices where people’s preferences tend to adhere to the resulting class or racial hierarchies (Kaplan and Illouz Reference Kaplan and Illouz2022), as the architectural features risk exacerbating these very inequalities. Conner (Reference Conner2023, 8) finds that “Grindr reproduces categories [according to physical attributes that align with the sociological literature on sexual hierarchies] by formatting their platform to emphasize those traits.” Another study finds a correlation between the frequent use of dating platforms and skepticism of multiculturalism (Callander, Holt, and Newman Reference Callander, Holt and Newman2012).

In sum, filters can threaten people’s access to sexual standing in several ways. By encouraging users to think about others as belonging to categories that one should opt out from interacting with, the app’s architecture will result in fewer potential matches for the filtered-away groups. Note that it does not need to directly discriminate against anyone to have this effect. By making it easy to make racist or ableist choices, filters can, as Sonu Bedi (Reference Bedi2019, 141; see also, Robinson Reference Robinson2015; Williams Reference Williams2024) argues, be seen as a form of more indirectly discriminatory, and equal-standing-undermining, steering of the kind the U.S. Civil Rights Act prohibits in other areas of society.

App architecture can also hide or make it harder to realize that one is making choices in line with social norms for desirability, or that one is considering only users who are similar to oneself. Tinder’s swipe function, for instance, encourages users to engage with the app in an absent-minded way, allowing “unconscious racial preferences to be expressed without troubling users’ perceptions of themselves as non-racist” (Narr Reference Narr2021, 3). When dating apps were first introduced, their creators could not be held responsible for people’s preferences. However, when we take seriously the vast impact these intermediaries have in the contemporary sexual sphere, we see that this is not true anymore.

Designers could choose differently in relation to users’ pre-existing preferences. As Bedi (Reference Bedi2019, ch4) argues about racial filters, removing them does not hinder anyone from acting on their racist preference. However, there is no duty on the part of dating app corporations to architecturally encourage—steer—users to opt out of considering potential matches belonging to certain social groups. In contrast, they owe all users not to undermine their claims of sexual standing and choice improvement.

To sum up, dating apps’ architectural choices have furthered people’s interests in noninterference and choice improvement. To better realize individuals’ sexual standing claims, they could do more to counteract patterns of discrimination through design. There are ways to, for instance, encourage users to explore outside the group of profiles it predicts the users would be most likely to “like” or say yes to (Hutson et al. Reference Hutson, Taft, Barocas and Levy2018, 10). This would amount to a transformative approach of the kind I discussed above.

MODERATION

Penalizing abusive users is one way dating apps can cater to others’ claims of noninterference in the sexual sphere. Through moderation, platforms exercise power by enforcing rules for conduct on the app, evaluating users’ reports of others’ behavior, protecting users from each other, settling disputes between users, and removing content that is harmful or illegal (Gillespie Reference Gillespie2018, 6; Williams Reference Williams2024, 149). As Tarleton Gillespie (Reference Gillespie2018, 5) argues, platforms come to:

When a user reports another, the app must weigh the users’ claims. If the app removes the reported, it will further the interests in noninterference of users who are similar to the reporter. At the same time, this makes the reported lose the opportunities to develop sexual standing and follow sexual life plans that exist on the app. Therefore, his claims about equal standing and choice improvement are undermined.

The potential harm on the side of the reporter includes psychological, physical, sexual, and economic harm. Over 50% of female users have been threatened, called offensive names, contacted despite declining, or sent unwanted explicit images (Pew Research Center 2023, 35–7). If the reported is harmful in his interactions with other users, not removing him would risk undermining the claims to noninterference of anyone the app matches him with. On the other hand, the reported risks being effectively blocked from meeting potential partners if the AI or human moderator decides to ban him. Dating apps now constitute crucial sites for developing sexual standing and following sexual life plans in the contemporary sexual sphere. If he is banned, his claims of sexual standing and choice improvement would be undermined. The stakes are high because blocking is often for life (Tinder N.d.).

Further, apps with the same owners are likely to use the same moderation practices over several big platforms, something that can be especially burdensome to users in the growing context of what Kleinberg and Raghavan (Reference Kleinberg and Raghavan2021) define as algorithmic monoculture, a situation in which many decision-makers use the same algorithms. Where several of the more prominent apps utilize identical systems or algorithms, this increases the risk for outcome homogenization, which is “the phenomenon of individuals (or groups) exclusively receiving negative outcomes from all decision-makers they interact with” in a particular sphere of life or society (Bommasani et al. Reference Bommasani, Creel, Kumar, Jurafsky and Liang2022, 2).

Unfortunately, there is reason to worry that moderation systems undermine innocent users’ claims to sexual standing and choice improvement, without at the same time furthering other users’ claims to noninterference. The report systems can be easy to misuse. For instance, some users employ “revenge report” strategies, reporting anyone who rejects them. Others use it to target trans people, people of color, members of other minoritised groups, and people who support specific political causes (Cheung Reference Cheung2022; Williams Reference Williams2024, 147). Unless the rules for moderation are adequately fine-tuned, the moderation AI can mistakenly be triggered by these report revengers to block users who have done nothing wrong. At the same time, it is notoriously difficult to complain about moderation decisions.

Overall, moderation systems must be made to better balance users’ claims of noninterference with claims of equal standing and choice improvement. While users have an interest in not being wrongfully interfered with, they also have an interest in being able to develop sexual standing and follow sexual life plans. There is a need for effective report systems that are harder to misuse for dubious or harmful purposes, and that are tailored to the multiplicity of users on these platforms. The experiences of users who are most at risk of harm, such as women, queer individuals, and members of racial minorities, must be centered in the development of such systems (Rigot Reference Rigot2022; Williams Reference Williams2024, 171–5).

In their communication with users, platforms must also be clear about how their moderation practices work, and about the steps taken when users make reports about, for instance, sexual harassment (Williams Reference Williams2024, 174). Without adequate moderation control, reporting risks becoming a weapon with which abusive users can punish people, rather than a tool with which abused users can protect themselves and others from unwanted interference (Larson Reference Larson2023). Since being blocked from these apps effectively hinders individuals from pursuing the kinds of lives they want to live, a justifiable moderation practice must provide access to due process systems in the case of wrongful banning.

AMPLIFICATION

Apps exercise power in a third way: through amplification. By increasing some content’s visibility and reach, at the expense of other content, the app determines who gets attention from whom (Keller Reference Keller2021). Based on information about users and their preferences, algorithms could, if adequately fine-tuned, predict which profiles a user is likely to match with. Therefore, this function offers a potential solution to the Matching Problem and a way for users to realize the claim of equal standing and choice improvement. Different users will be benefited or penalized depending on the apps’ use of specific algorithms—processes or rule sets “by which calculations are made, problems solved, or decisions reached” (Fischli Reference Fischli2022a, 132; see also Hao Reference Hao2018). Even though it is up to the individual user whether, when presented with a potential match, they will accept or reject it, the algorithm provides the options in the opportunity set. Users are presented with options, but they are also presented to others as options.

Recall that apps cannot help but intervene—they need to sort profiles in some order. The option to do nothing does not exist. Without amplification, the app would not help users find people they are likely to share sexual life plans with or in whose eyes they have sexual standing. Such an app would be a poor solution to the Matching Problem. It also would not necessarily be fairer to “do nothing” than to sort profiles according to some other rule. For instance, if by “do nothing” we mean chronological sorting by when people signed up to the app, this would arbitrarily favor users who have been members for longer. It is not obvious we should think this is a better way to distribute opportunities in the sexual sphere than some other algorithmic rule.

A user who is presented as an option to few others has fewer opportunities to develop sexual standing. Filters are therefore not the only feature that disadvantage certain groups of users. To solve the matching problem, the amplification algorithm sorts users according to perceived features, often through built-in assumptions about connections between physical similarity and compatibility: as Apryl Williams (Reference Williams2024, 3) shows, the dating industry’s “decisions about whom you might be attracted to (and whom you may attract) are largely influenced by how you look, how attractive the algorithm deems you to be, and how often other highly attractive individuals have interacted with your profile.” However, the algorithm need not be designed with this purpose to have the same effect. It will treat users who do not have certain preferences as if they do have them, if they share certain features with those who have the preferences in question. When many people reject users with specific characteristics, the algorithm learns to see similar profiles as less attractive. In turn, this gives these profiles fewer matches (Nader Reference Nader2020). This process can, as Narr (Reference Narr2021, 6, see also Williams Reference Williams2024, ch2) points out in the case of racism, “make it impossible to circumvent the racial biases of the entire network of past and present users.” The problem is aggravated if the app uses these users’ data to predict the preferences of new users whom the app codes as similar (Wang et al. Reference Wang, Liu, He, Jiang, Du, Tang, King, Chen and Wang2011), especially as this process will not be visible to the affected users.

Dating apps know that amplification affects users’ self-respect and how to monitor this effect. For instance, the Tinder Boost pay feature is presented to users as a means to increase their chances of getting matches. It “amplifies a profile’s visibility for half an hour in the physical area where the user is currently located. The profile is put on top of others” “stack of recommendations without indicating this to these other users” (Courtois and Timmermans Reference Courtois and Timmermans2018, 7). As the Swedish Tinder user Simon, 23, explains,

In the data economy, ‘there exists an entire ecosystem designed to capitalize on individuals” vulnerabilities’ (Fischli Reference Fischli2022a, 141). As the dating app operates in the sexual sphere, where we are particularly likely to be vulnerable, this has especially troubling effects. Emil, a man in his thirties, tells the newspaper Svenska Dagbladet:

With the help of its data collection, the dating app has ample resources to predict when users are “at their lowest” and most likely to buy to get that boost—and their self-esteem back. A recent report shows that Tinder charged some customers up to 12 times more for the same service than others, without neither alerting users to this, nor providing an explanation as to why prices differed (Sveriges Konsumenter Reference Konsumenter2022). Because Tinder collects data about gender, sexual orientation, age, and geographic area, they have the means to discriminate against certain users. In response to the report, Tinder promised that prices were not discriminating according to certain features but simultaneously refused to reveal how they set their prices on Tinder (Eklund Reference Eklund2022). Following a dialogue with the European Commission, the company later committed to no longer use personalized pricing without clearly informing users (European Commission 2024). Note that no one is interfering with users like Emil and Simon. However, through their amplification decisions, some apps have made it unreasonably costly for people like them to realise their claims of sexual standing and choice improvement.

Unfortunately, despite their potential to provide more predictable, efficient, and objective results than human decision-making, algorithms often incorporate the biases of their creators. Bias in algorithms can often be explained by the fact that developers share our societies’ assumptions and that they train their algorithms on the data of people living in these societies (Goggin, Ellis, and Hawkins Reference Goggin, Ellis and Hawkins2019; Lillywhite and Wolbring Reference Lillywhite and Wolbring2019; Tilmes Reference Tilmes2022).

As the Tinder case shows, we do not need to know an algorithm’s rules to have our behavior effectively constrained and ruled by it (Gillespie Reference Gillespie, Gillespie, Boczkowski and Foot2014). When algorithms shape our behavior, it is harder for users to pin down the content of the rule shaping us and where it comes from. This feature is an essential source of the dating platforms’ power to dominate users. The inequality between the masses, who provide the data, and those who collect and analyze them, is steadily growing, and the only people who understand the expensive and complicated technologies, datasets, and software that impact so much of our lives, are those who are in control over them (Andrejevic Reference Andrejevic2014, 1676; Fischli Reference Fischli2022a, 131–4; Spiekermann et al. Reference Spiekermann, Slavni, Axelsen and Lawford-Smith2021; Williams Reference Williams2024, 170–1).

While dating apps clearly can use amplification to provide people with valuable opportunities to follow a wide range of sexual life plans, these opportunities are presently only made available to those who accept the dating app corporations’ unfavorable terms. The only way to resist the power of this system is to log off and meet up with potential partners in some other forum. But in a compartmentalized society where most people meet potential partners through dating apps, this option is less viable—especially for sexual minorities for whom the Matching Problem is especially salient.

Philosophers have shown how, in the digital public sphere, users struggle to hold platforms accountable because these exercise power in a nontransparent way (Aytac Reference Aytac2022, 9; Everett Reference Everett2018; Flew and Wilding Reference Flew and Wilding2021). This problem is arguably amplified in the context of sex and intimacy, where political organizing, consciousness-raising, and activism are less palatable due to the stigma surrounding sex and the resulting risk of distress—few would want to stand under the banner of the “unmatched.”Footnote ²⁰

THE LIMITS OF MARKETS

Before we return to the question of whether dating app power can be justified, we must turn to an objection. Given that individuals freely download dating apps, readers might think this consent already constitutes a justification for their power. They might wonder why, beyond basic legal requirements, we should consider any duties on the part of individual apps. Could we not simply trust the market to ensure people’s interests are realized?Footnote ²¹

There are a few things to say in response. First, as Lazar (Reference Lazar2023; Reference LazarForthcoming) argues concerning the digital public sphere, for consent to platforms to have the justifying effect, there needs to exist sufficiently good alternatives to being involved in the digital market. As dating apps grow more pervasive, leaving the digital sexual sphere becomes costlier and less realistic. As we saw above, the institution of dating has evolved alongside the rise of dating app corporations, making its goods and roles difficult to access without being affiliated with them.

However, the free-market objector might respond, even if there are no nonmarket alternatives, there might be sufficiently good alternatives within the market. If this were the case, dating apps would give us less cause for concern. Discontented users could simply leave any app that does not meet the requirements of justice and join one that better serves their interests, creating incentives for tech companies to design apps that do realize them.Footnote ²² In other words, one could argue that there are circumstances where voluntary actions between consenting adults should be limited, but these are cases where such limitation is necessary to realize some social good. If the market is enough to realize the good, subjecting actors to further demands is wrong.

While this objection might theoretically be valid in ideal settings, it does not hold in practice. Simply put, the market has thus far not generated apps that fully realize individuals’ claims. This is unsurprising; under the mode of production that characterizes the digital economy, it is unlikely that the market will provide apps that do not prioritize data gathering and extraction over everything else—this, after all, constitutes a main source of profit (Cohen Reference Cohen2019).Footnote ²³ Even though there are different apps, the market of dating apps is characterized by “mimetic isomorphism”: whether niche or mainstream, apps tend to copy competitors, resulting in product homogenization. Consequently, this market has failed to generate meaningful diversity when it comes to platform features (Bergström Reference Bergström2022, 40–2, 47–8).Footnote ²⁴ Further, network effects make the bigger apps hard to leave; to be worthwhile, a dating app must reach a certain level of users. Even the most ethically governed app would be useless without enough potential partners to interact with.

Finally, limits on markets are justified when the voluntary actions of individuals harm others, for instance, through negative externalities for third parties (Heath Reference Heath2006). Dating apps have such externalities. The online dating industry has changed what it means to have sex, to initiate intimacy, and how we relate to each other both as partners and potential partners (Bergström Reference Bergström2022, ch4; Illouz Reference Illouz2007, 90; Reference Illouz2012, 180–3). As we saw, apps’ design choices about architecture, moderation, and amplification not only affect crucial practices in the sexual sphere but aggravate, for example, racial stereotyping. Apps shape what life is like even for those who would never date online (compare, Lazar Reference Lazar2023; Reference Lazarforthcoming).

So far, I have responded to the free market objection on its terms: I showed that even though individuals freely sign up and consent to apps’ terms, it would be justified to limit some aspects of the contemporary market in dating apps. However, it is important to emphasize that I resist the objection’s fundamental premise, namely that the normative landscape of intimacy initiation, including the operations of dating apps, is best described as a set of voluntaristic actions by consenting adults. One main aim of this paper is to show exactly the opposite: intimacy initiation is a social institution impacting society and individual lives. Since this is the case, we cannot reach relevant normative conclusions by primarily focusing on specific individual actions or wishes. Taking this seriously amounts to seeing that, like with other social institutions, there is justification for limiting voluntary actions, especially of powerful actors. By intervening in and shaping the institution, apps acquire responsibility to ensure that its goods and roles are distributed fairly, rather than simply in a way that keeps users online. The existence of different apps does not change this. Instead, dating apps are collectively responsible for providing a digital sexual sphere where people’s interests are not undermined. The responsibility will be greater for actors that make up a greater share of the market, but the assessment of an individual’s opportunities in the digital sexual sphere must include not only the functioning of the particular app she is using but also the full flora of available apps. What matters is whether a user can realize her claims; if she lacks opportunities on one app but has opportunities on the rest, she is unlikely to have a complaint.

JUSTIFYING DATING APPS

Going back to the question of whether the power of dating apps can be justified, we are now able to say that these intermediaries do offer individuals something that they, as a requirement of justice, are owed. Dating apps’ ability to provide users with something valuable—opportunities to develop sexual standing and form, revise, and pursue sexual life plans—is the most plausible explanation of the value of online dating available to liberal egalitarians. Note, however, that we only see this if we recognize that we have something like claims to equal standing and choice improvement in the sexual sphere.

Dating apps offer solutions to the three major problems of the traditional, nondigital sexual sphere. Since dating apps connect people to potential partners and provide a system where users can reveal their intent and preferences vis-à-vis particular dates and intimate relationships generally, they are particularly beneficial for sexual minorities. They have provided access to opportunities to develop sexual standing in the eyes of relevant others and follow their sexual life plans to groups who have been worse off in terms of such opportunities. In terms of liberal egalitarian justice, this improvement in the choice situations of people who have had and continuously have the weightiest complaints about the sexual sphere is something that significantly counts in favor of dating apps.

However, dating apps only offer such opportunities partially, and to some individuals. Thus, they are not doing enough to realize the claims we have in the sexual sphere. There is reason to think that the apps’ design features and practices threaten some people’s claims to noninterference, equal standing, and choice improvement. These apps also disempower people by making them dependent on unexplained and arbitrary high-stake decisions. The dating apps’ lack of publicity concerning algorithms and data processing contributes to the power imbalance between companies and users that has arisen after the digital revolutions. It should make us pressure tech companies to be transparent about what is “under the hood”: to know that and how an algorithm is being used (Reich, Sahami, and Weinstein Reference Reich, Sahami and Weinstein2021, 105). One way to ensure that app companies’ use of user data and algorithms is in line with our values and interests would be to make them provide explanations of how they balance these demands (Adadi and Berrada Reference Adadi and Berrada2018; Suzor et al. Reference Suzor, West, Quodling and York2019; Vredenburgh Reference Vredenburgh2022).Footnote ²⁵

While, as noted above, a systematic treatment of non-digital-sphere solutions falls outside the immediate scope of this paper, there is reason to underline that the state also must enable people to meet outside of the digital sexual sphere. This would counteract the compartmentalization of the sexual sphere, while also resisting the development by which tech companies colonize and benefit from private areas of our lives. Promoting free time, supporting reform of unjust social norms, and providing open and safe public spaces, inclusive sexual education as well as a functioning nightlife, are some of the measures the state can implement to improve the realization of individual claims in the sexual sphere.

In the meantime, dating app workers have reason to contribute to the establishment of a professional organization that can monitor the actions and impact of its members. Reich, Sahami, and Weinstein (Reference Reich, Sahami and Weinstein2021, 246–52) have proposed a “Hippocratic oath for tech” setting up ethical guidelines for this technologists’ organization which would be able to license workers and sanction misconduct. While dating app developers would fall under a general professional organization for tech, their practice is special. Governing the digital sexual sphere comes with particular responsibilities. Yet, the producers of dating apps have often lacked expertise regarding social practices relating to intimate relationships. As one of Bergström’s (Reference Bergström2022, 42–3) dating app technologist respondents puts it, “I made a carpooling site that doesn’t work that well, I have the know-how to do websites, why wouldn’t I make a dating site? […] The advantage is that technically it’s the same thing.”

Acknowledging dating apps’ role in realizing social justice in the sexual sphere, an important area for further research and public deliberation is to what extent developers of systems for sexual matchmaking should be subject to sexual-sphere-specific rules. It seems reasonable to expect that such matchmaking, based on its shared features with sexology and therapeutic work, requires skills that go beyond those of the general tech worker. We might also suspect that if dating apps were developed by psychologists and sociologists rather than by professionals trained primarily in technology and computer science, they would look quite different. At the very least, there is reason for the dating app organization to implement a voluntary code of conduct, on which moderation, amplification, and architectural choices should take into account everyone’s claims to noninterference, equal standing, and choice improvement in the sexual sphere. This could, for instance, involve features that encourage users to be more open and creative in their choice of potential partners.

Regrettably, the tech sector has thus far failed to self-regulate. There is therefore a case for intervening in the market of dating platforms to bring their conduct more in line with the demands of social justice for the sexual sphere. Tech companies have adjusted their business in Europe in response to legislation such as GDPR, which, as Williams (Reference Williams2024, 176) argues, shows that when given the incentive, they can increase transparency. Exactly how this regulation should look is a question for further research and something that would need to be subjected to consultation with and deliberation among the public, democratic representatives, dating app companies, and other stakeholders.

Since dating apps are matchmaking tools, it appears that any rules or potential regulations must be symmetric and apply to anyone who professionally matches romantic mates. One argument against this has to do with scale—matchmaking a few individuals at a time and matchmaking millions by automatic means may be relevantly different. This is an issue that needs more attention, and I am here only able to tentatively say it is plausible to think that my analysis applies to all kinds of matchmaking. One reason that matchmaking has not previously been regulated may be that sexual standing and sexual choice have not been taken seriously as concerns for social justice. It is time that was changed.