← Return to search results
Back to Prindle Institute

The Ethics of Manipulinks

image of computer screen covered in pup-up ads

Let’s say you go onto a website to find the perfect new item for your Dolly Parton-themed home office. A pop-up appears asking you to sign up for the website’s newsletter to get informed about all your decorating needs. You go to click out of the pop-up, only to find that the decline text reads “No, I hate good décor.”

What you’ve just encountered is called a manipulink, and it’s designed to drive engagement by making the user feel bad for doing certain actions. Manipulinks can undermine user trust and are often part of other dark patterns that try to trick users into doing something that they wouldn’t otherwise want to do.

While these practices can undermine user trust and hurt brand loyalty over time, the ethical problems of manipulinks go beyond making the user feel bad and hurting the company’s bottom line.

The core problem is that the user is being manipulated in a way that is morally suspect. But is all user manipulation bad? And what are the core ethical problems that manipulinks raise?

To answer these questions, I will draw on Marcia Baron’s view of manipulation, which lays out different kinds of manipulation and identifies when manipulation is morally problematic. Not all manipulation is bad, but when manipulation goes wrong, it can reflect “either a failure to view others as rational beings, or an impatience over the nuisance of having to treat them as rational – and as equals.”

On Baron’s view, there are roughly three types of manipulation.

Type 1 involves lying to or otherwise deceiving the person being manipulated. The manipulator will often try to hide the fact that they are lying. For example, a website might try to conceal the fact that, by purchasing an item and failing to remove a discount, the user is also signing up for a subscription service that will cost them more over time.

Type 2 manipulation tries to pressure the person being manipulated into doing what the manipulator wants, often transparently. This kind of manipulation could be achieved by providing an incentive that is hard to resist, threatening to do something like ending a friendship, inducing guilt trips or other emotional reactions, or wearing others down through complaining or other means.

Our initial example seems to be an instance of this kind, as the decline text is meant to make the user feel guilty or uncomfortable with clicking the link, even though that emotion isn’t warranted. If the same website or app were to have continual pop-ups that required the user to click out of them until they subscribed or paid money to the website, that could also count as a kind of pressuring or an attempt to wear the user down (I’m looking at you, Candy Crush).

Type 3 manipulation involves trying to get the person to reconceptualize something by emphasizing certain things and de-emphasizing others to serve the manipulator’s ends. This kind of manipulation wants the person being manipulated to see something in a different light.

For example, the manipulink text that reads “No, I hate good décor” tries to get the user to see their action of declining the newsletter as an action that declines good taste as well. Or, a website might mess with text size, so that the sale price is emphasized and the shipping cost is deemphasized to get the user to think about what a deal they are getting. As both examples show, the different types of manipulation can intersect with each other—the first a mix of Types 2 and 3, the second a mix of Types 1 and 3.

These different kinds of manipulation do not have to be intentional. Sometimes user manipulation may just be a product of bad design, perhaps because there were unintentional consequences of a design that was supposed to accomplish another function or perhaps because someone configured a page incorrectly.

But often these strategies of manipulation occur across different aspects of a platform in a concerted effort to get users to do what the manipulator wants. In the worst cases, the users are being used.

In these worst-case scenarios, the problem seems to be exactly as Baron describes, as the users are not treated as rational beings with the ability to make informed choices but instead as fodder for increased metrics, whether that be increased sales, clicks, loyalty program signups, or otherwise. We can contrast this with a more ethical model that places the user’s needs and autonomy first and then constructs a platform that will best serve those needs. Instead of tricking or pressuring the user to increase brand metrics, designers will try to meet user needs first, which if done well, will naturally drive engagement.

What is interesting about this user-first approach is that it does not necessarily reduce to considerations of autonomy.

A user’s interests and needs can’t be collapsed into the ability to make any choices on the platform that they want without interference. Sometimes it might be good to manipulate the user for their own good.

For example, a website might prompt a user to think twice before posting something mean to prevent widespread bullying. Even though this pop-up inhibits the user’s initial choice and nudges them to do something different, it is intended to act in the best interest of both the user posting and the other users who might encounter that post. This tactic seems to fall into the third type of manipulation, or getting the person to reconceptualize, and it is a good example of manipulation that helps the user and appears to be morally good.

Of course, paternalism in the interest of the user can go too far in removing user choice, but limited manipulation that helps the user to make the decisions that they will ultimately be happy with seems to be a good thing. One way that companies can avoid problematic paternalism is by involving users at different stages of the design process to ensure that user needs are being met. What is important here is to treat users as co-deliberators in the process of developing platforms to best meet user needs, taking all users into account.

If the user finds that they are being carefully thought about and considered in a way that takes their interests into account, they will return that goodwill in kind. This is not just good business practice; it is good ethical practice.

No Fun and Games

image of retro "Level Up" screen

You may not have heard the term, but you’ve probably encountered gamification of one form or another several times today already.

‘Gamification’ refers to the process of embedding game-like elements into non-game contexts to increase motivation or make activities more interesting or gratifying. Game-like elements include attainable goals, rules dictating how the goal can be achieved, and feedback mechanisms that track progress.

For example, Duolingo is a program that gamifies the process of purposefully learning a language. Users are given language lessons and tested on their progress, just like students in a classroom. But these ordinary learning strategies are scaffolded by attainable goals, real-time feedback mechanisms (like points and progress bars), and rewards, making the experience of learning on Duolingo feel like a game. For instance, someone learning Spanish might be presented with the goal of identifying 10 consecutive clothing words, where their progress is tracked in real-time by a visible progress bar, and success is awarded with a colorful congratulation from a digital owl. Duolingo is motivating because it gives users concrete, achievable goals and allows users to track progress towards those goals in real time.

Gamification is not limited to learning programs. Thanks to advocates who tout the motivational power of games, increasingly large portions of our lives are becoming gamified, from online discourse to the workplace to dating.

As with most powerful tools, we should be mindful about how we allow gamification to infiltrate our lives. I will mention three potential downsides.

One issue is that particular gamification elements can function to directly undermine the original purpose of an activity. An example is the Snapstreak feature on Snapchat. Snapchat is a gamified application that enables users to share (often fun) photographs with friends. While gamification on Snapchat generally enhances the fun of the application, certain gamification elements, such as Snapstreaks, tend to do the opposite. Snapstreaks are visible records, accompanied by an emoji, of how many days in a row two users have exchanged photographs. Many users feel compelled to maintain Snapstreaks even when they don’t have any interesting content to share. To achieve this, users laboriously send meaningless content (e.g., a completely black photograph) to all those with whom they have existing Snapstreaks, day after day. The Snapstreak feature has, for users like this, transformed Snapchat into a chore. This benefits the company that owns Snapchat by increasing user engagement. But it undermines the fun.

Relatedly, sometimes an entire gamification structure threatens to erode the quality of an activity by changing the goals or values pursued in an activity. For example, some have argued that the gamification of discourse on Twitter undermines the quality of that discourse by altering people’s conversational aims. Healthy public discourse in a liberal society will include diverse interlocutors with diverse conversational aims such as pursuing truth, persuading others, and promoting empathy. This motivational diversity is good because it fosters diverse conversational approaches and content. (By analogy, think about the difference between, on the one hand, the conversation you might find at a party with people from many different backgrounds who have many different interests and, on the other hand, the one-dimensional conversation you might find at a party where everyone wants to talk about a job they all share). Yet Twitter and similar platforms turn discourse into something like a game, where the goal is to accumulate as many Likes, Followers, and Retweets as possible. As more people adopt this gamified aim as their primary conversational aim, the discursive community becomes increasingly motivationally homogeneous, and consequently the discourse becomes less dynamic. This is especially so given that getting Likes and so forth is a relatively simple conversational aim, which is often best achieved by making a contribution that immediately appeals to the lowest common denominator. Thus, gamifying discourse can reduce its quality. And more generally, gamification of an activity can undermine its value.

Third, some worry that gamification designed to improve our lives can sometimes actually inhibit our flourishing. Many gamification applications, such as Habitify and Nike Run Club, promise to help users develop new activities, habits, and skills. For example, Nike Run Club motivates users to become better runners. The application tracks users across various metrics such as distance and speed. Users can win virtual trophies, compete with other users, and so forth. These gamification mechanisms motivate users to develop new running habits. Plausibly, though, human flourishing is not just a matter of performing worthwhile activities. It also requires that one is motivated to perform those activities for the right sorts of reasons and that these activities are an expression of worthwhile character traits like perseverance. Applications like Nike Run Club invite users to think about worthwhile activities and good habits as a means of checking externally-imposed boxes. Yet intuitively this is a suboptimal motivation. Someone who wakes up before dawn to go on a run primarily because they reflectively endorse running as a worthwhile activity and have the willpower to act on their considered judgment is more closely approximating an ideal of human flourishing than someone who does the same thing primarily because they want to obtain a badge produced by the Nike marketing department. The underlying thought is that we should be intentional not just about what sort of life we want to live but also how we go about creating that life. The easiest way to develop an activity, habit, or skill is not always the best way if we want to live autonomously and excellently.

These are by no means the only worries about gamification, but they are sufficient to establish the point that gamification is not always and unequivocally good.

The upshot, I think, is that we should be thoughtful about when and how we allow our lives to be gamified in order to ensure that gamification serves rather than undermines our interests. When we encounter gamification, we might ask ourselves the following questions:

    1. Is getting caught up in these gamified aims consistent with the value or point of the relevant activity?
    2. Does getting caught up in this form of gamification change me in desirable or undesirable ways?

Let’s apply these questions to Tinder as a test case.

Tinder is a dating application that matches users who signal mutual interest in one another. Users create a profile that includes a picture and a short autobiographical blurb. Users are then presented with profiles of other users and have the option of either signaling interest (by swiping right) or lack thereof (by swiping left). Users who signal mutual interest have the opportunity to chat directly through the application.

Tinder invites users to think of the dating process as a game where the goals include evaluating others and accumulating as many matches (or right swipes) as possible. This is by design.

“We always saw Tinder, the interface, as a game,” Tinder’s co-founder, Sean Read, said in a 2014 Time interview. “Nobody joins Tinder because they’re looking for something,” he explained. “They join because they want to have fun. It doesn’t even matter if you match because swiping is so fun.”

The tendency to think of dating as a game is not new (think about the term “scoring”). But Tinder changes the game since Tinder’s gamified goals can be achieved without meaningful human interaction. Does getting caught up in these aims undermine the activity of dating? Arguably it does, if the point of dating is to engage in meaningful human interaction of one kind or another. Does getting caught up in Tinder’s gamification change users in desirable or undesirable ways? Well, that depends on the user. But someone who is motivated to spend hours a day thumbing through superficial dating profiles is probably not in this respect approximating an ideal of human flourishing. Yet this is a tendency that Tinder encourages.

There is a real worry that when we ask the above questions (and others like them), we will discover that many gamification systems that appear to benefit us actually work against our interests. This is why it pays to be mindful about how gamification is applied.

Why Would Anyone Marry?

black and white photograph of "Just Married" in back window of vintage car

“Marriage is wonderful when it lasts forever, and I envy the old couples in When Harry Met Sally who reminisce tearfully about the day they met 50 years before. I no longer believe, however, that a marriage is a failure if it doesn’t last forever. It may be a tragedy, but it is not necessarily a failure. And when a marriage does last forever with love alive, it is a miracle.”

                                                                           —Peggy O’Mara, Mothering

On May 3rd, Bill and Melinda Gates announced their divorce, stating that after lots of “thought and a lot of work on our relationship, we have made the decision to end our marriage,” ending their twenty-seven year union. Whatever their reasons for divorce — the announcement being understandably vague — we should pause here to think about an important philosophical question, namely: why would anyone marry? Choosing to enter into a marriage is among the most consequential decisions one can make, affecting potential future children, well-being and happiness, and so on. A good marriage can be a major blessing, but a bad marriage can be hell.

Here’s a rough argument against marriage: no one wants a loveless marriage; it is something that, if we knew it would happen, would likely encourage us not to wed to begin with — to be stuck in a loveless marriage is a potential horror show. Often enough, the love couples feel for each other, entering a marriage, will fade and sometimes cease. To avoid this fate, we should be reticent prior to taking the vows. As the philosopher, Dan Moller, explains:

“Reduced to a crude sketch, the argument [against marriage] is simply that, (a) most of us view the prospect of being married in the absence of mutual love with something like horror or at least great antipathy; (b) the mutual love between us and our spouse existing at the inception of our marriage may very well fail to persist; and hence (c) when we marry we are putting ourselves in the position of quite possibly ending up in a loveless marriage of the sort we acknowledge to be undesirable.”

What’s partly hard about marrying is that emotions — an essential aspect of romantic love — have an autonomous element, not wholly governed by the will, even with the best of intentions. If romantic love is an essential aspect of marriage, how can marriage vows bind? When you cannot control something, you cannot be culpable for it — we don’t think people culpable for the ocean waves, by example.

Someone may object that marriage vows specify actions, not emotional states, and are thus under our control. As the philosopher, Justin McBrayer, argues:

“Notice how heavily [many generic marriage vows focus] on actions compared to emotions: support one’s partner, honor one’s partner, respect one’s partner, and so on. Even the emotional content is easily understood in a behavioral sense: to be a faithful partner in sickness and health clearly has a behavioral component. To see this, imagine the following thought-experiment. Suppose Landon makes the aforementioned promise to Hannah. Suppose next that he feels all the right things toward her (for example, he is in love with her), but that his behavior is wildly erratic – he sleeps around, is verbally abusive to Hannah, abandons her when she is ill, etc. Would anyone be willing to say that Landon has fulfilled his wedding vow? Surely not.”

There’s only one problem with this: it ignores the fact that emotions, like romantic love, are also essential to marriage. Of course it is clear that marriage vows, and marriage itself, includes the vow to do certain things — many married people expect their partner to be faithful to them, by example. However, it cannot be that actions and behavior is all there is to marriage — without an appropriate emotional component, it is not clear one would be in a marriage they would find fulfilling and satisfying. Imagine a different thought experiment: John treats his wife well, cares for her when ill, never cheats on her, and so forth, but simply doesn’t feel love for her at all. He is a good husband to fulfill his marriage vows because his nagging conscience won’t let him break them. We wouldn’t think though that this is a fulfilling marriage; it likely isn’t one John would have entered into had he known the result would be the death of love for his wife. Whether marriage vows include a behavior aspect is irrelevant — emotions seem a necessary part of marriage too.

So then we must wonder why people would marry — the people who marry usually do not want to end in a loveless marriage. There must be an upside then to marriage; benefits that contribute to our well-being. Likely they partly include the positive emotions many experience, especially in the early days of the marriage. However, there must be more to explain why so many people, despite the risk a bad marriage poses to living a good life, still choose to tie the knot.

It looks like, in addition to a romantic venture, marriage is a commitment device: by imposing costs on dissolving a marriage, the institution of marriage forces individuals to comprise and grow in a manner they wouldn’t if the cost of ‘walking away’ from a relationship were relatively low. And this can do several things: incentivize individuals in a marriage to work together, to engage in personal improvement for the sake of the union and family, provide a more stable environment for rearing kids, and make relationship-specific investments. Think of it like this: often time, but not always, couples will get along if forced to. Imagine, by example, you’re permanently handcuffed to someone — your fate and theirs are bound up. It would then make sense, assuming the other person is reasonable, to make the best of the situation by doing things like getting along and compromising.

This can, of course, be a bad thing if one of the partners is unreasonable; there are downsides to nearly anything. Our point isn’t to claim that marriage is always a good thing — we began the piece by thinking about why anyone would marry — but to highlight some marital benefits; whether the benefits outweigh the costs will likely vary from couple to couple. However, there are some downsides too if the cost of abandoning commitments is too low: it may be harder in some cases to find a long-term relationship if the cost of finding a new partner, whenever one is even remotely dissatisfied, is too low. To illustrate, consider a scene from Season 5, Episode 4 of Seinfeld, where Elaine has just dumped her partner:

JERRY: You’re out of your mind you know that.

ELAINE: What?

JERRY: It’s an exclamation point! It’s a line with a dot under it.

ELAINE: Well, I felt a call for one.

JERRY: A call for one, you know I thought I’ve heard everything. I’ve never heard a relationship being affected by a punctuation.

ELAINE: I found it very troubling that he didn’t use one.

We don’t want to be trapped in a loveless marriage, obviously; but we should worry too if the cost of abandoning our commitments is too low — the ramification of that would extend beyond marriage to, among other moral practices, promises more generally.

“Nudges” and the Environmental Influences on our Morals

A photo of a telephone booth

This article has a set of discussion questions tailored for classroom use. Click here to download them. To see a full list of articles with discussion questions and other resources, visit our “Educational Resources” page.


Richard Thaler, a behavioral economist, won the Nobel Prize in economics this year. He co-authored the book, Nudge, in 2008. The theory behind “nudges” (a term he coined) changed the perspective of economics on the agents to be studied. Instead of picturing humans as rational preference satisfiers, Thaler suggests that we are susceptible to all sorts of irrational pressures and rarely do we decide to behave in ways that can be modeled on principles of rationality and our individual preferences. The “nudge” is one tool he uses in order to see one way in which we deviate from the rationalistic model of classical economics.

A “nudge” is a non-rational psychological factor that makes it more likely that someone behave in one way rather than another. Thaler himself came up with the model for behavior when he became frustrated at the speed with which he and his roommates were going through their stash of cashews. They took familiar enough means to remedy the problem – they kept the cashews out of the living rooms so that they’d have to trek to the kitchen to fill up, or started to hide them. Dieters everywhere have come up with similar solutions to reduce their intake of their favorite food – putting chips into a portioned bowl instead of eating from the bag, for instance. Taking steps like these are providing yourself with “nudges” to eat less: doing these things makes it more likely that you will exhibit self-control.

We can see intuitively why this might be. By putting food into portions, you need to exert effort to refill, and by putting the food at a distance away that also ups the effort level. Along with this extra effort expenditure comes a break in behavior: now you can’t more or less “mindlessly” continue to consume. You face a choice when your bowl is empty: get up and refill? Do you really want or need more? So there is more chance you will stop. You’re helping yourself out with these nudges.

Along these lines, we are more likely to opt for healthy food if it is at eye level when we are hungry – otherwise we’ll go for our normal junk food of choice. The arrangement of grocery stores can have an impact on the healthy choices of its customers by arranging food accordingly and making it easier to choose healthful options.

Thaler noted how irrational we are as agents. We make decisions mostly based on convenience and speed, and fall victim to irrational decision structures like sunk costs.

Perhaps the most famously effective nudges have been outside the realm of food: the presence of an insect on the bottom of a urinal raises the likelihood that people urinating make their target. There are also policies spreading worldwide that have citizens opt out of organ donation programs if they prefer not to be donors rather than opt in if they would like to be a donor. This change to policy increased participation significantly, despite the fact that according to classical economics, the method of selection should be a neutral factor: the number of people who want to be donors should end up being donors either way.

In Chicago in 2016, the ideas24 anti-gun program used the underlying psychology of Thaler’s views to develop an approach to working with incarcerated teenagers. Noting that people behave in a scripted way when under stress, teenagers were encouraged to note triggers and rewrite the scripts, with “group lessons around decision making” that take Thaler’s views to heart by focusing on things that would nudge them away from violence: “In one lesson, inmates list the people who may be affected by something they’ve done.” Nudge theories are on the horizon for use in decreasing gun crime in New York City as well as other cities in the United States. Daniel Webster, director of Johns Hopkins Center for Gun Policy Research, hopes that nudging people to believe that carrying a gun isn’t normal – “that their behavior is ‘out there’” – will affect change.

The key commitment behind implementing nudges in order to affect change in behavior is that the way that individuals decide what to do isn’t on the basis of deciding what they think is best and behaving accordingly. The nudge is in place based on an external determination about what is best to do – external to the agent’s decision-making process at the time. This has led to some controversy in politics: to what extent should the government be actively trying to shape the behavior of its citizens?

Recall the benighted soda tax in New York City. Though cities such as Philadelphia and Berkeley have a tax on sugary drinks, intended to dissuade people from purchasing excessive amounts of the detrimental beverages, when such a bill was proposed in New York City, it was met with public outrage.

We can ask questions beyond the legitimacy of such government policies. If our behavior can be so easily influenced by factors outside of our control, or disconnected from our preferences and commitments about what to do, then how will that affect our notions about how responsible we are for our actions?

The impact of nudges on our behavior fits with a family of psychological studies in the second half of the twentieth century that showed the significant impact of apparently irrelevant features of the environment on our moral behavior.

Intuitively, it is important to develop and cultivate a good moral personality or character. We care about having good character traits ourselves and look for good traits in one another. Traits such as honesty, compassion, bravery, humility, etc. are desirable and typically relevant for assessing the behavior we come across in the world. If we take someone to be honest, we think they will be more likely to tell the truth. This is why we value having honest friends – we take them to be more reliable in this regard. If someone we take to be honest lies on one occasion, the fact that we take them to have this trait of honesty typically allows us to chalk the lie up to a rare or one-time event. We can lean on the reliability of the trait and maintain the relationship.

The research on nudges may break the connection between character traits and behavior. For when a nudge results in a behavior, something other than the character trait was the cause of the behavior. This is in line with other psychological studies that showed trait-irrelevant factors to be better predictors of behavior than these traits that our moral practices tend to favor.

Consider a sampling of the well-known studies. The “dime experiment,” conducted in 1972 by Isen and Levin, tested the likelihood that subjects would help someone who dropped a large amount of papers on the street (the paper-dropper being a participant in the experiment). The subjects had just used a payphone, and those that found a dime left in the phone were significantly more likely to help than those that did not find a dime. In what we could call the “smell experiments,” people were more fair and generous when in a room that smelled clean. Finally, the proximity of a perceived authority figure significantly affected subjects’ willingness to cause pain to a stranger in the famous Milgram experiments. These effects did not track the moral makeup of the agents in question, yet predicted their moral behavior.

The effectiveness of nudges, and the success of the manipulation of subjects in the experiments sampled above, suggests a worry for our practices of moral evaluation. If someone acts in a certain way because of a smell, or because they found a dime, or because of a nudge, to what extent was the action their own? Would we blame or praise them to the same extent if they had acted without the external factor?

The more we learn about human agency, the more we must accept that we are influenced by a host of external factors. Thaler’s work to a large degree suggested harnessing this feature for good – using the influence-ability of our agency to direct it towards higher-order goals. This may be the direction our blame and praise practices head towards as well: how well we are managing our agency, rather than individual actions. On such a picture, we are creatures to manage and direct, responsible for driving ourselves well.