Mobile Devices and Autonomy: Individual-Level Effects

Abstract

In this chapter, we offer a review of the empirical findings that animate our concerns about the effects of mobile devices on individuals (the troublesome findings about the effects of mobile devices on collectives is reviewed in Chap. 7). We begin by clarifying what we mean by mobile devices, noting that these devices are not our only concern. We also explain why they are our primary concern. In short, we are especially concerned by mobile devices because their mobility, combined with their ability to track our behavior and provide real-time feedback to it, make them powerful behavior modification devices. After explaining how mobile devices can modify our behavior and why—given the structure of the attention economy—the interests of the developers of these technologies are not always in alignment with our own, we show that there is reason to think that mobile devices have detrimental effects on all eight facets of autonomy outlined in the preceding chapter.

For getting other human beings’ inclinations into one’s power, so that one can direct and determine them according to one’s intentions, is almost the same as possessing others as mere tools of one’s will.

—Kant, Anth. 7:271

The thought process that went into building these applications … was all about: “How do we consume as much of your time and conscious attention as possible?” And that means that we need to sort of give you a little dopamine hit every once in a while, because someone liked or commented on a photo or a post or whatever. And that’s going to get you to contribute more content, and that’s going to get you … more likes and comments. It’s a social-validation feedback loop … exactly the kind of thing that a hacker like myself would come up with, because you’re exploiting a vulnerability in human psychology.

—Sean Parker, first president of Facebook (This quote from Parker comes from an interview with The Guardian. See Solon (2017a)).

3.1 Introduction

In this chapter, we lay important groundwork for our coming arguments. We clarify which technologies we are chiefly concerned with, as well as the empirical grounds for thinking that they conflict with our (individual-level) autonomy.¹ This, we will argue, gives rise to a duty we have to ourselves to be “digital minimalists,” that is, to be mindful about our interactions with digital technologies such that they do not conflict with our capacity to set and pursue our ends. But we will not complete that argument—or flesh out the concept of digital minimalism—until the next chapter. This chapter is a clearinghouse for the empirical evidence for the argument for digital minimalism, which also serves as a foundation that we will build upon to justify the empirical premises needed for the arguments for attention ecology (a duty to assist others in being digital minimalists) and group-level digital minimalism (the duty we have, as a collective, to engage in digital minimalism).

Throughout the chapter, it will be useful to draw on Esther’s example from the introduction. Esther joked about how difficult it is for her to focus on reading a book as she finds herself compulsively unlocking her phone to look at Instagram. Of course, we do not think that Esther’s experience is exhaustive; the concerns about problematic smartphone use go well beyond issues with literacy.² But it will be a useful lens for exploring some of the concerns that arise in the empirical literature.

3.2 What Are the Devices?

Our main concern in this book is our relationship with what we will call “mobile devices,” in the context of the attention economy and a culture of constant exposure to mobile devices.

We will clarify the term and explain why mobile devices are our main concern, but first note that many devices, namely desktop computers, do not qualify as mobile devices. This does not mean that our concerns do not extend to these non-mobile devices. We are also concerned with non-mobile, web-connected screens in the context of the attention economy; but this is not our main concern. Now, for clarifications.

By “mobile device” we do not simply mean devices that are portable. If we did, then mechanical watches would qualify. But we are not concerned with people’s relationships with mechanical watches. Instead, we mean handheld computers that have access to the internet, such as smartphones, tablets, and even laptops (though, as we explain in what follows, smartphones are of utmost concern). These devices allow you to be in constant contact with the attention economy, even when you are outside the home and while you are in bed.

Let us now explain why we think the context—the attention economy and a culture of constant contact with it—is important.

We’ll begin by explaining what we take the attention economy to be and why we do not take it on its own (or at least without further contextualization) to be an object of concern. Following Castro and Pham (2020), we take the attention economy to be the economic market where consumers give media developers their attention in exchange for a service (e.g., a news feed), and where developers sell consumer attention to advertisers.

We do not take the attention economy alone to raise the sorts of concerns that animate us here. As Castro and Pham (2020) note, the sorts of exchanges that constitute the attention economy at least go back to the 1830s, when The New York Sun transitioned away from subscriptions to advertising as its main source of revenue.³ To go from the existence of the attention economy to our concerns about autonomy, further elements are needed. In our case those elements are behavior modification devices that can habituate us to having our attention stolen at any moment (i.e., mobile devices), and the normalization of the constant use of these devices. We would now like to say a bit more about these other elements.

When reflecting on the relationships we have with our mobile devices—especially in the context of the attention economy—it is important to consider two important facts about them. First, they are designed to get our attention. They can ring, vibrate, light up, and display notifications. This is for good reason: we do not want to miss important calls, texts, or reminders. Second, these features can be used to instill habits in us. Some of these habitual behaviors appear to be compulsive (and to be products of manipulation).⁴

Consider Nir Eyal’s bestselling design guide Hooked: How to Build Habit-Forming Products, which opens by boasting about how much we use our phones:

Seventy-nine percent of smartphone owners check their device within fifteen minutes of waking up … one-third of Americans say they would rather give up sex than lose their cell phones. … people check their phones … an astounding 150 [times a day]. Face it: We’re hooked. (Eyal 2014, 1)

And what is it, according to Eyal, that keeps us hooked? It’s the power of these devices to create and foster habits, which he identifies as “‘automatic behaviors triggered by situational cues’: things we do with little or no conscious thought” (Ibid.).

Eyal sees behavior modification via the use of mobile devices as an important opportunity for companies:

Forming habits is imperative for the survival of many products. As infinite distractions compete for our attention, companies are learning to master novel tactics to stay relevant in users’ minds. (Ibid. 2)

He then goes on to introduce the centerpiece of his book, the Hooked Model: “a four-phase process … to form habits” (Ibid. 8).

The Hooked Model is simple but powerful. As he says, it’s what drives us to touch our phone first thing in the morning and then over a hundred times after that; what makes a third of us prefer our phones to sex; what lies behind, “the pull to visit YouTube, Facebook, or Twitter” that keeps you “tapping and scrolling an hour later” even if you just wanted to use one of those services “just a few minutes” (Ibid. 1).

The four phases of the Hooked Model run as follows.

First, trigger the user. The idea here is to cue the (potential) user to perform an action (“action” being the next step). This can be done via a notification (e.g., “Someone has visited your profile”) or prompt (“To Continue Reading, Register for Free”). As Eyal notes, triggers can be both external (e.g., notifications, prompts) or internal, which happen when “a product becomes tightly coupled with a thought, and emotion, or a preexisting routine” (Ibid. 47). One of the goals is to use the Hooked Model so effectively that users develop internal triggers, “the brass ring of habit-forming technology” (Ibid. 48).

Second, if the first phase went as planned, the user will perform an action, or “behavior done in anticipation of a reward” (Ibid. 7). Following up on the examples above, this could include checking your profile or creating an account. The keys to triggers that manifest actions, Eyal notes, are two factors as they relate to action: the underlying motivation to perform the action and the ease of performing it. The suggestion is to increase motivation and ease. He notes, “To initiate action, doing must be easier than thinking” (Ibid. 61).

Third, the action is to be associated with a variable reward, that is, the action is to be associated with an unpredictable possibility of reward:

Research shows that levels of the neurotransmitter dopamine surge when the brain is expecting a reward … which suppresses the areas of the brain associated with judgment and reason while activating the parts associated with wanting and desire. (Ibid. 9)

Eyal goes on to cite slot machines and lotteries as classic examples of habit-forming products that effectively use variable rewards. He then goes on to give a quick fictionalized example of a Pinterest user, who, in the presence of interesting and uninteresting photos, has her brain dopamine system set “aflutter with the promise of reward”; before she knows it, she has “spent forty-five minutes scrolling” (Ibid. 11).

The fourth, and final, phase of the Hooked Model involves the user making an investment, that is, having them put something of value into the product, something that takes time or effort. The idea here is to prime the user so that they are more susceptible to the next trigger sent their way.

As Eyal notes, citing work by Dan Ariely, Michael Norton, and Daniel Mochon, “we irrationally value our efforts” (Ibid. 136). He goes on to say that this, along with a drive to seek consistency in our behaviors and an aversion to cognitive dissonance, can be leveraged to bring users back. This means, Eyal states, that investments can get users to rationalize their use of the product, speeding up the formation of habits: “Rationalization helps us give reasons for our behaviors, even when those reasons might have been designed by others” (Ibid. 141).

An important element of all of this is the fact that it is normal to have the device on your person at all times. Unlike its predecessors—such as the desktop computer (which can run the Hooked cycles, but doesn’t fit in your pocket) or print newspapers (whose content could be subsidized through advertising but couldn’t run Hooked cycles)—a mobile device can go anywhere we might want to go, making us constantly available for the operant conditioning that Eyal promotes in his book. And as Fabio et al. (2022) note in their study on smartphone use and behavioral and cognitive self-control deficits, “the fact that a smartphone is small, easily handled and portable makes the risks more insidious and pervasive” (2).

Why, then, are we not worried about desktop computers? Again, we are, at least in some cases (such as the problems related to social media, which we discuss in Chap. 7), but they are not our primary concern. People could have damaging relationships with their desktops, absent mobile devices and the attention economy (e.g., gaming or gambling addictions). However, we do not see this as central to the widespread concerns we have about autonomy, as the constant connection, feedback, and monitoring made possible by mobile devices are what create the conditions for the scaled attention economy that worry us.

3.3 How Mobile Devices Undermine Autonomy

We will now discuss the individual-level effects of mobile devices, in terms of the two facets of autonomy discussed in Chap. 2 (viz. capacity and authenticity). We should note that these do not constitute the only empirically oriented concerns of the book; they are concerns that primarily play a role in our first major argument of three: our argument for the (individual-level) duty to be digital minimalists. We discuss group-level issues (such as polarization and democracy) in Chap. 7.

We should also note that we have some humility about the results discussed below (and throughout the book). What is discussed below is a snapshot as of this writing of our understanding of the effects of mobile devices on users. While we are confident in our analysis, we should note that studying these effects is complex and often muddled by, for example, the fact that it is difficult (if not impossible) in many cases to find a control group for the study. Furthermore, this is a burgeoning area of psychological and neuroscientific research, so the data are fairly new, and research is ongoing. Nevertheless, we will present the scientific findings that are currently available, and we cite large-scale meta-analyses whenever possible.

3.3.1 Capacity

Recall from the last chapter that one source of autonomy deficits is effects on the capacities required for autonomy: baseline capacities, freedom from external constraints, absence of cognitive inhibitions, and having a sufficiently wide range of options. We begin with the effects of mobile devices on these capacities. In this section, we discuss reasons to think that mobile devices (or our relationships with them) have negative effects on all of them. Having a problematic relationship with your phone has several negative effects on your baseline capacities. As we explain in the next section, it affects executive function, working memory, cognitive functioning, sleep, and much more.

3.3.1.1 Baseline Capacities

We begin with effects on baseline capacities (the ability to form intentions, make coherent judgments, exercise executive function, etc.). These findings can be sorted into two kinds. We can call one of these direct effects, those which report a direct association between mobile device use and some negative outcome for autonomy, such as the inability to pay attention. The others are indirect, they report on an association between mobile device use and some outcome that isn’t itself negative for autonomy as such, but which is associated with a negative outcome. Here, loss of sleep is a good example: loss of sleep itself isn’t a diminution of a capacity, but we do know that the underslept tend to have diminished capacities.⁵

Following the meta-analysis by Wilmer et al. (2017), we will organize our discussion of direct effects into the following categories: attention, memory, delay of gratification, and general cognitive functioning. While these capacities might not be a complete list of the basic capacities essential for autonomy, they must at least come close. In order to set ends, we need to be able to attend to reasons that speak in favor of or against a choice. In order to pursue those ends, we need to monitor our progress and reconnoiter as obstacles come and go.⁶ We need to remember what we have done or what we want to do next. We need to resist temptations that will ultimately set us back, and we need to avoid mundane errors.

Let us begin with a discussion of attention. One of the main concerns in this literature is the impact of mobile devices on focused attention, “the ability to maintain a directed attentional focus over an extended period of time” (Ibid. 4). In their discussion of attention, Wilmer et al. (2017) make the helpful distinction between exogenous and endogenous intrusions on attention. These at least roughly correspond to what Eyal calls “external” and “internal” triggers, with the former—for Wilmer et al. (2017)—occurring when an environmental cue (e.g., a notification) fuels an interruption and the latter occurring when “[one’s] own thoughts drift toward a smartphone-related activity, and thereby evince an otherwise unsolicited drive to begin interacting with the device” (Ibid. 4).

As Wilmer et al. (2017) note, both sorts of interruptions can initiate a chain of other distractions:

Once attention has been shifted to the smartphone for one purpose (e.g., by virtue of a specific notification source), users often then engage in a chain of subsequent task-unrelated acts on the smartphone, thereby extending the period of disruption. Studies exploring these “within-phone” interruptions have found that task completion in one app can be delayed by up to 400% by an unintended interruption from another app (Leiva et al. 2012). And, some evidence suggests that the more “rich” (e.g., including a visual image rather than just text) the information encountered during an interruption, the more detrimental the distraction is likely to be with respect to primary task completion. (Levy et al. 2016). (Ibid. 4)

These are the sorts of interruptions that the Hooked Model aims to make us more susceptible to, and they are emblematic of the sorts of interruptions that concern us from the point of view of autonomy, because, among other things, these are the sorts of interruptions that introduce errors into our thinking and execution of tasks we have set for ourselves. Consider, for instance, resumption errors, “errors that arise in task performance that is resumed following an interruption or task-switch” (Ibid. 5).⁷ The likelihood of committing such an error increases sharply when an interruption exceeds 15 seconds, which is a threshold that smartphone interruptions—which often snowball in the way described above—commonly exceed.⁸ We are even more susceptible to these sorts of endogenous interruptions out of a desire for stimulation when our ongoing tasks fail to entertain us.⁹

Turning now to exogenous interruptions, it is striking how disruptive mobile devices can be. We noted earlier that mobile devices are often engineered to grab attention, with alarms, lights, and vibrations. It is worth noting, however, that we seem to be so conditioned to give these devices our attention that their mere presence distracts us and introduces error even when they are off and even when they are not our own device.¹⁰ Screens themselves seem to leech attention. Exposure to notifications—let alone responding to them—of course exacerbates these effects. Indeed, it has been shown that exposure to notifications “significantly damages performance on an attention-demanding task” (Stothart et al. 2015, 896). A meta-analysis by Caird et al. (2014) confirms this concern, and they demonstrate the troubling implications for driving, where the introduction of error can be fatal.

Now we will turn from attention, which selects which information to focus on, to working memory, which keeps information readily available for processing. The mere presence of a mobile device has a detrimental effect on available working memory.¹¹ Ward et al. (2017) demonstrate that the strength of this effect is moderated by how much one uses their phone: “Ironically, the more consumers depend on their smartphones, the more they seem to suffer from their presence—or, more optimistically, the more they may stand to benefit from their absence” (149).

Not only does the presence of a phone impact working memory, behaviors associated with smartphones have negative impacts on working memory as well. For instance, a number of studies show that media multitasking (e.g., scrolling through a feed on one’s phone while streaming a show on one’s laptop) is linked to diminished working memory,¹² both when external distractions are absent or present.¹³

Taking stock of what we have said so far, it seems, at the very least, that the presence of phones—even if they are turned off—interferes with attention and working memory, two essential features of reasoning to a conclusion about what to do and then doing it. Let us now take a look at another important factor in achieving one’s goals: delay of gratification.

In their (2016) paper, Wilmer and Chein demonstrate evidence of an oft-hypothesized connection between mobile device use and delay of gratification. Specifically, they find a significant correlation between device usage and one’s discount rate of future rewards, that is, their susceptibility to immediate gratification. Now, as Wilmer et al. (2017) note, like many studies, it is correlational: it could, on the basis of these findings, just be that mobile device use is evidence of higher discount rates. However, there is at least one experimental result that suggests that the causal arrow flows in the direction that worries critics of mobile devices. Hadar et al. (2015) ran a longitudinal study where participants lacking smartphone experience were tested before and after receiving smartphones. Subjects experienced an increase in impulsivity and a decrease in information processing three months after receiving the devices.

Now that we have reviewed some of the studies that speak to the corrosive effect that mobile devices have on attention, memory, and delay of gratification, let’s look at the effects of mobile device use on cognitive functioning more generally.

One way to get a sense of the impact mobile devices have here is by looking at their effect on grades.¹⁴ We can begin by noting that there is a large number of studies showing a negative correlation between mobile device use and academic performance.¹⁵ For instance, Lee et al. (2017) and Dietz and Henrich (2014) show, via random assignments of treatment conditions, that there is a causal connection between mobile device use and the comprehension of material from a lecture. It is also interesting to note that one of these studies—Sana et al. (2013)—showed how laptop use in class not only affects the user but all those who could see the screen (Sana et al. 2013, cf. Wilmer et al. (2017)). This shows how the negative effects of mobile devices can flow beyond the individual user. It is not just your cognition that is affected but also those around you. There are also reasons to believe that mobile device use affects cognition outside the context of education. A study from Baumgartner et al. (2014) demonstrates a correlation between multimedia multitasking and failures of executive function. Cain et al. (2016) corroborated these links in a laboratory setting (cf. Wilmer et al. (2017)).

All of these effects of smartphones have a direct impact on capacities linked to autonomy (executive function, working memory, attention span, etc.). But smartphone use also has an indirect effect on our baseline capacities. As we have noted, this happens in the case of sleep deprivation. Although sleep is not one of the capacities that is linked to autonomy, sleep deprivation leads to a variety of cognitive deficits. We will close this section with some brief remarks about the effects mobile devices might have on sleep, anxiety, and depression—factors that have an impact on cognitive functioning.

It is fairly well-known that mobile device use is associated with lack of sleep and that lack of sleep has negative impacts on cognition (Lim and Dinges 2008, cf. Wilmer et al. 2017). For instance, in their meta-analysis of the association between nighttime mobile device use and sleep outcome, Carter et al. (2016) found “a strong and consistent association between bedtime media device use and inadequate sleep quantity … poor sleep quality … and excessive daytime sleepiness” (1203). Further, companies have an obvious interest in keeping us awake, expanding the pool of attention from which they can draw to show us advertising or to sell us products. The profit motive is the driving factor behind this feature of the attention economy.¹⁶ Companies maximize ad revenue by maximizing engagement. The more time we spend looking at screens, the more ads we see. As Johann Hari recounts in Stolen Focus from his conversation with physician and sleep researcher Charles Czeisler:

In a society dominated by the values of consumer capitalism, “sleep is a big problem” … “If you’re asleep, you’re not spending money, so you’re not consuming anything. You’re not producing any products.” He explained that “during the last recession [2008] … they talked about global output going down by so many percent, and consumption going down. But if everybody were to spend [an] extra hour sleeping [as they did in the past], they wouldn’t be on Amazon. They wouldn’t be buying things.” If we went back to sleeping a healthy amount … Charles said, “it would be an earthquake for our economic system, because our economic system has become dependent on sleep-depriving people. The attentional failures are just roadkill. That’s just the cost of doing business.” (Hari 2022, 76–77)

It is no wonder that mobile devices conflict with sleep. Most of us keep our phones at arm’s reach while we sleep, and it can be difficult to resist the pull of the attention economy as it entices us with its triggers to stay up a little longer to consume more content.

Moving on to anxiety and depression, which are themselves associated with diminished cognitive functioning (American Psychiatric Association 2013, cf. Wilmer et al. 2017), a number of recent studies have shown a causal connection between mobile device use and negative mental health outcomes. For instance, after monitoring 143 undergraduates for a week, Hunt et al. (2018) randomly assigned severe limits to social media to some undergraduates but not others for three weeks. During the three weeks, the “limited use group showed significant reductions in loneliness, depression, anxiety, and fear of missing out” (Ibid.). Similar causal results have been reported by Sherlock and Wagstaff (2019), Lambert et al. (2022), and Brailovskaia et al. (2022).

These results should not surprise us. In her breakout book How to Do Nothing, Jenny Odell describes what she, in unplugging from the attention economy, wants to escape:

To me, one of the most troubling ways social media has been used in recent years is to foment waves of hysteria and fear, both by news media and by users themselves. Whipped into a permanent state of frenzy, people create and subject themselves to news cycles, complaining of anxiety at the same time that they check back ever more diligently. The logic of advertising and clicks dictates the media experience, which is exploitative by design. Media companies trying to keep up with each other create a kind of “arms race” of urgency that abuses our attention and leaves us no time to think. (Odell 2019, 59)

Once this “logic of advertising and clicks” is laid bare, it comes as no surprise that, as Hunt et al. (2018) and others have found, social media makes us anxious, depressed, and worse off in other ways. Thus, it is no surprise that mobile device exposure can lead to a diminution of the basic capacities that are necessary for autonomy.

3.3.1.2 External Constraints

Given the negative effects of mobile devices and the attention economy, the idea of unplugging entirely might sound appealing. But research has found that this is harder to accomplish than you might think. As we will explain later in our discussion of authenticity, people routinely report that they use their devices more than they would like to. One survey found that 62% of US consumers have made efforts to cut back on their smartphone usage, but only half of them had any success.¹⁷

Indeed, smartphone and social media users often report feeling “addicted” to their devices, and you would have to pay them to quit. As we mentioned in the introduction, a group of economists conducted one of the largest studies of the effects of social media on well-being in 2018. They recruited a sample of 2743 Facebook users, and the researchers offered them various amounts of money to quit using Facebook for four weeks. They wanted to determine how much money it would take to get someone to quit for just one month. The result ended up being higher than most estimates. The mean figure was $180.¹⁸ Ironically, even though users had to be paid to quit Facebook, most of them were happier without it. Allcott et al. (2020) found that those who quit for a month reported increased well-being equal to the jump one would get from earning about $30,000 more in annual income (654). During the experiment, they spent more time with family and friends, and they used Facebook much less after the experiment was over.

But in many cases, our relationships with our phones and social media are not built on our own terms. As we saw in the last section, many of us are trapped in the cycles the Hooked Model runs us through, which were deliberately developed so that companies could maximize the amount of time we spend looking at screens. What is more, we are often subject to external constraints that make it so that unplugging is not a live option. In the last chapter, this was presented as the second component of capacities that are required for autonomy. This idea is fairly straightforward. The more you are subject to external constraints, the less autonomous you are. When it comes to our use of mobile devices, perhaps the most obvious instance of external constraints is employment.

It is not uncommon for employers to require employees to be in constant communication by email or through other digital channels (e.g., Slack, text message, WhatsApp). Under these conditions, an employee is never truly off the clock, and this all but necessitates an unhealthy relationship with one’s smartphone. Cal Newport (2021b) argues that these working conditions predictably lead to burnout and high turnover. He suggests that these policies are not in employers’ best interests as they undermine productivity rather than promote it. But he is particularly concerned about the effect this has on employees’ well-being. When employers require “slavish devotion to in-boxes and chat channels, then this adds up to a whole lot of global miserableness! From a utilitarian perspective, this level of suffering cannot be ignored—especially if there is something that we might be able to do to alleviate it” (Newport 2021a).¹⁹

Workers often find that they have no choice but to comply with the technological demands of their employers. If a worker is highly skilled and is able to find employment elsewhere, then she may have global autonomy with respect to this issue, but employees typically lack local autonomy in these matters. To fight back against this situation, workers in Europe have won what they call “the right to disconnect.” In 2016, this right became part of French labor law; companies with more than 50 employees are now required to negotiate after-hours email policies with employees or their unions. This law was introduced after a report found that many workers were using their work phones at home, and the majority of them wanted more control over their use of technology.²⁰

But the majority of workers have much less control over their employer’s technology policies. And this is one way that we are subject to external constraints when it comes to our relationships with mobile devices. The situation in schools is quite similar, as the use of devices like tablets and laptops has become a mandatory part of the curriculum for most students. A survey conducted by the US Department of Education found that 94% of public schools used them for the 2022–2023 school year.²¹

We have witnessed this firsthand in the lives of our own children. The Miami-Dade public school system requires students to use electronic devices starting at age 4. Like 10 million other students in the US, children in Miami-Dade are required to use a digital curriculum called “i-Ready.”²² We were dismayed (but not surprised) to learn that i-Ready includes a reading activity designed to look like Instagram. When a five-year-old student correctly answers a question on “Yoop-o-Gram,” she is rewarded with a flurry of likes and hearts.

To be clear, we are not simply reactionaries who are opposed to the use of all technology in school. As we will explain in the next chapter, we are not suggesting that we should become Luddites and eschew technology altogether; when used and designed properly, mobile devices can enhance our autonomy. Obviously, students should learn how to use technology, and schools would fail to promote students’ autonomy if they did not introduce them to a variety of digital tools. So we are not advocating for the abolition of digital curricula. Our point here (as elsewhere) is simply that we should be thoughtful about how we integrate technology into our lives, and this is especially pressing when the use of technology is mandatory. It is easy for educational administrators to see the upsides of using technology in school. When used appropriately alongside classroom instruction from teachers, digital tools can be indeed useful ways of pursuing learning outcomes. But there are good reasons to be careful about harnessing the addictive components of these devices.

Gamification—the idea behind the hearts and likes on Yoop-o-gram—often looks like a great way to increase engagement, but it has downsides as well. Thi Nguyen (2020, 2021) argues that gamification is alluring insofar as it purports to be a way of getting people to do otherwise boring and repetitive tasks. By pursuing points, people might be more willing to perform those tasks. What makes games appealing, in his view, is that they offer “value clarity.” In many domains of our lives, it is very difficult to know which values to promote at which times. Should you be working, cleaning your house, or doing something else entirely? But by simplifying things to a matter of points, games offer clarity. The goal becomes clear and straightforward (e.g., get more hearts, get more likes).

When this is applied to other domains of life (in the form of gamification), it has the effect of flattening out our values. Perhaps students should be interested in education for a variety of reasons: personal growth, edification, enhanced autonomy, etc. But when education is gamified (e.g., through the single-minded pursuit of maximizing GPA), all other values disappear as we become victims of what he calls “value capture.” Value capture involves a radical simplification of a wide range of values as it flattens them into something quantifiable. This can have perverse effects. Students who were once interested in learning and taking on difficult challenges might start taking less challenging, “easy A” courses in pursuit of a high GPA.²³

This point about gamification has important implications for the use of technology in schools. When it comes to digital technology, we should recall McLuhan’s claim that “the medium is the message” (McLuhan and Lapham 1994). In many cases, the content on the smartphone is not the entire story. The fact that the content is on the smartphone (or tablet) is also relevant. When it comes to promoting literacy in children, researchers are unequivocal about the importance of getting children to appreciate and enjoy reading.²⁴ This often requires that books be present in the child’s home, that she sees adults in her life reading books, etc. As neuroscientist and literacy advocate Maryanne Wolf has shown, getting children to become interested in books has become more difficult in a context where they must compete with smartphones, tablets, and YouTube.²⁵ She explains how reading involves an amazing interplay of neural components—vision, cognition, affect, attention, and language. Reading awakens both hemispheres and all five layers of the brain. Given that our brains did not evolve for reading, acquiring this ability requires neuroplasticity. The brain rewires itself to make deep reading possible. The danger, of course, is that we can undo this as well. So although we see lots of words on screens (news stories, social media posts, etc.), there is a vast neurological difference between this kind of shallow reading and deep reading.²⁶

Neuroscientists have argued that this effect may be especially pronounced in “digital natives” who grew up with internet technologies. Loh and Kanai (2014) write:

With multifaceted affordances, the Internet environment has profoundly transformed our thoughts and behaviors. Growing up with Internet technologies, “Digital Natives” gravitate toward “shallow” information processing behaviors characterized by rapid attention shifting and reduced deliberations. They engage in increased multitasking behaviors that are linked to increased distractibility and poor executive control abilities. Digital natives also exhibit higher prevalence of Internet-related addictive behaviors that reflect altered reward-processing and self-control mechanisms. Recent neuroimaging investigations have suggested associations between these Internet-related cognitive impacts and structural changes in the brain. (506)

For this reason, we may want to be cautious about how we use technology in schools, especially with young children. It is tempting to think that we are enhancing children’s ability to read by having them sound out words on a tablet, but there are reasons to worry that the child might learn to sound out words but never develop a desire to read books (whether physical or electronic). Perhaps we should not be surprised when children grow up to find themselves in the same situation as Esther Povitsky. Much as they might like to read a book, they will find it difficult to resist the urge to put it down to look at Instagram or TikTok. In some instances, external constraints limit our options. Some children have no choice but to learn to read on screens.

There are many cases where we are compelled to adopt the ends that others have set for us. We often have to accept the terms that our schools or employers give us. In and of itself, this can undermine autonomy. Given that autonomy involves setting and pursuing your own ends, this capacity is surely undermined when someone else sets the terms of your life for you. What is more, when external constraints require us to have unhealthy relationships with mobile devices, this can have downstream effects in other areas of our lives. As we saw in the previous section, it can weaken some of our baseline capacities. In the next section, we will explore other negative effects, including the ways that mobile devices and social media undermine our self-esteem.

3.3.1.3 Cognitive Inhibitions

The quote from Sean Parker at the beginning of this chapter notes that we are often brought to our mobile devices via the exploitation of vulnerabilities in human psychology, including the drive for social validation. As former Google design ethicist Tristan Harris notes, “Everyone innately responds to social approval … That’s why it’s so important to recognize how powerful designers are when they exploit this vulnerability” (Harris 2016). Here, we explore one consequence of the power of exploiting this vulnerability: its effects on self-esteem. As we will soon show, there is a tight connection between certain mobile device applications and low self-esteem (i.e., being critical of oneself, downplaying one’s positive qualities, judging oneself as inferior to others, using negative words to describe oneself, assuming luck plays a large role in one’s successes, blaming oneself when things go wrong, and disbelief in compliments about oneself). In other words, interactions with mobile devices seem to generate a host of cognitive inhibitions of the sort that we think are detrimental to autonomy. Much like Benson’s example of the wife in Gaslight, we come to question our own capacities in such a way that we rely on others and forfeit some of our agency to them.

Most of us are familiar with the idea that news feeds, especially on image-driven apps like Instagram, present a distorted—if not entirely false—image of others’ lives. Indeed, some lives—such as that of Shudu²⁷ (244,000 followers at the time of writing) and Lil Miquela (3 million followers at the time of writing)—are complete fabrications, entirely computer generated. They are, as one reporter put it, “physically perfect women made of pixels, standing in for women who have long been pressured to become physically perfect” (Tiffany 2019).

While many users are aware of the fact that what we see on Instagram is unrepresentative of real life, this doesn’t stop us from comparing ourselves or each other to them. Indeed, a comment on a photo of Lil Miquela posing with a human model reads, “the robot more pretty [sic].”²⁸ Meanwhile, many other comments express confusion over whether Lil Miquela is a “robot.” In a post promoting a makeup brand, Shudu is complimented on her beauty and told that her eyeshadow works well with her skin tone. She is asked, “how do I become a model [?]”²⁹ These comments seem to assume that Shudu is a human. This is understandable. Shudu’s representations are photorealistic and easily mistaken for actual photos. Further, the company she is promoting in the post seems—as far as we can tell—to exclusively sell real, physical makeup intended for use on flesh and blood humans.

These are no doubt extreme examples, but the point here generalizes beyond extreme cases. Despite the fact that we know that posts on sites like Instagram are idealizations, when we are exposed to these images they nevertheless enter into our social comparisons. These comparisons include upwards social comparison, where we compare ourselves to others that we think are faring better than ourselves (Festinger 1954). Indeed, one study—Fardouly and Holland (2018)—exposed women ages 18 to 25 years old to posts from social media containing images of attractive women, with some participants but not others being shown the images with a disclaimer, such as, “Not real life—I didn’t pay for this outfit, took countless photos trying to look hot for Instagram” (4317), and others being shown the images without the disclaimers. Women in both groups left the exposure feeling much worse about their bodies than those in the control group, who were shown images of travel. This suggests that even if we are aware of the distortions of social media, we engage in detrimental upwards social comparisons anyways.

With this in mind, it should come as little surprise that there seems to be a fairly robust association between social media use and low self-esteem. For instance, Kelly et al. (2019) consulted data from 10,904 English 14-year-olds and found that increased “social media use was associated with, among other things, low self-esteem and body image,” with this effect being larger for girls than boys (60). Twenge and Farley (2021) corroborate these results, and add to our understanding by showing that the effects not only vary by gender, but also by activity: “Hours spent on social media and Internet use were more strongly associated with self-harm behaviors, depressive symptoms, low life satisfaction, and low self-esteem than hours spent on electronic gaming and TV watching” (207).³⁰

The findings by Twenge and Farley might raise a challenge to our project: given that it isn’t mobile device exposure as such, but what one does on the mobile device that matters, the problem here could be said to lie with social media, as opposed to mobile devices. A few notes are in order here. First, as noted above, mobile devices aren’t our only concern; they are simply our primary concern. Our larger concerns have to do with the effects of web-connected screens on autonomy, but, as we have claimed above, there are reasons to focus on a more narrow target than this, given the power that mobile devices have, in particular, to influence behavior.

Further, we believe that our focus on mobile devices—due to their portability and powers of influence—is only reinforced by the negative effects of social media. Here it is worth noting that most social media access now occurs via phones. Indeed, recent estimates based on Facebook’s own data show that more than 98% of users use their phone to access the site, with over 80% exclusively accessing the site via phone.³¹ The fact that access via phone is so high is not an accident: that we can check in on our social networks anywhere we might be (in bed, at school) in otherwise free moments (going to the bathroom, waiting at a stoplight), and that this checking can become a habit (in Eyal’s sense), allows us to be exposed to this material more than we would otherwise be able to.

What we have said so far might give the impression that social media use only has an effect on the self-esteem of young teenagers. But this is not so. For instance, Hanna et al. (2017), in a study of undergraduate men and women, found a positive correlation between Facebook use and self-objectification (how much one is “preoccupied with how their body appears to others” 174), social comparison (“comparing oneself with others” 172), depression, anxiety, and low self-esteem. Several studies mirror Kelly et al. (2019) in finding distinct gender patterns, but in adults. For instance, Miljeteig and von Soest (2022) found that while social media was correlated with stability of self-esteem among both men and women, they also found that, for women, recent low self-esteem was more predictive of current social media use. They further found that social media use was more predictive of current low self-esteem. With this they note that their findings “support the notion of a reciprocal relationship between social media use and self-esteem for women, where self-esteem level may motivate women to use social media more frequently and social media may be a source of lower self-esteem” (373).

Up until now, most of the studies that we have presented in this section have been correlational. As always, this relationship could just be a result of the fact that lower self-esteem drives users to social media. However, there are studies that show that the causal direction flows in the other direction. Ozimek and Bierhoff (2020), for instance, exposed some subjects but not others to Facebook, with the subjects exposed to Facebook reporting lower self-esteem post exposure. Vogel et al. (2014) found similar results. Sherlock and Wagstaff (2019) found that exposure to beauty and fitness Instagram images significantly reduced self-assessments of attractiveness. Finally, Wolfe and Yakabovits (2022) were able to show that undergraduate women exposed to posts containing edited photos were more likely to edit their own photos when asked to then take a selfie. The same study also showed that “photo editing was associated with adverse changes in perceived attractiveness and mood” (1).

The takeaway here is that mobile devices expose us to content that could instill the sorts of cognitive inhibitions that put a drag on our autonomy. This is not entirely accidental. As we have seen through the testimony of people like Eyal, Parker, and Harris, the architects of services designed to run on our phones—and thus to go with us everywhere we go—will leverage vulnerabilities such as our inborn desire for social validation to get our eyes on the screen. However, in seeking this validation we often engage in activities that lower our opinions of ourselves due to our natural tendency to compare ourselves to those that we take to be faring better than ourselves. This is harmful to us in many ways. Diminished self-esteem negatively affects many aspects of well-being, and, more germane to our present concern, it can be disastrous for personal autonomy.

3.3.1.4 Options

In articulating the importance of options, we gave the example of a child, who, while formally free to do something else, considered being a coal miner as his only live option. No one literally forces the child into coal mining; no one denies him other options. But, he chooses as though he has no other options. This highlights the importance of not only having other options but being alive to them, being able to see them as options.

In On Education, Harry Brighouse argues for this point forcefully, on the grounds that it is an important facet of autonomy. He concludes that “children have a right to learn about a range of ways of living and to the kind of education that will enable them to reflect on their own way of life in the light of these alternatives, and, ultimately, to revise or reject the way of life their parents would pass down to them” (Brighouse 2006, 2). As Brighouse stresses, this is the only way for children to make informed decisions about which lives work for them. Citing Mill approvingly, he stresses the importance of exposing a child to those with convictions that differ from those of her parents and teachers:

Nor is it enough that he should hear the arguments of adversaries from his own teachers, presented as they state them, and accompanied by what they offer as refutations. That is not the way to do justice to the arguments, or bring them into real contact with his own mind. He must be able to hear them from persons who actually believe them; who defend them in earnest and do their very utmost for them. (Mill 1988, 35)

With this in mind, let us ask whether children—or we, ourselves—have sufficient access to less connected lives, to see if we can see them as suitable for us.

We suspect that we do not, and that this is not entirely due to the presence of external constraints mentioned above (though, of course, external constraints only exacerbate the concern). Indeed, despite study after study showing, e.g., that we are happier when we have less access to our email (Mark et al. 2012) or Facebook (Allcott et al. 2020), most of us, in the face of that knowledge, turn back to our devices.

On this last front, Ezra Klein shares an observation that, for us, resonates because we are so familiar with the phenomenon:

There’s a state I get in, less and less these days, but in part because of the way my world works and my phone and my computers. I now associate it with plane flights because nobody can call me, and I don’t buy internet. It’s a state that I only seem to access when reading, and only when reading without distraction for a long period of time. It’s very strange, and it is one of my most loved states. (Klein 2022)

What fuels Klein’s ability to access this state in the air? His hypothesis—and ours—is freedom from the internet, freedom from a fully functional smartphone. Yet what does he—what do we—do when he lands? Deliver himself back into the clutches of his phone:

[E]very time I get off of a plane, I say to myself, I’m going to do that more … I’m going to sit, and I’m going to have quiet time with a book … And then I don’t. (Ibid.)

And why? To be sure, there are many reasons (Hooked cycles, habits, the chaos of ordinary life, and so on). But among them there is the fact that it is simply, for many, unimaginable to live life without a phone.

This fact isn’t one that only middle-aged journalists and professors are familiar with. As “Billie,” one of the teenage girls Nancy Jo Sales interviewed for her book American Girls: Social Media and the Secret Lives of Teenagers, says, teens often feel that they do not have an option to not have a smartphone: “You have to have an iPhone. It’s like Apple has a monopoly on adolescence” (Sales 2017, 251; cf. Castro and Pham 2020). Billie is not alone in feeling this way. “Emily,” a teen Jean Twenge interviewed for her book iGen, hits a similar note: “Everyone uses it. It’s a good way to, like, make plans with people. If you don’t, you might miss out on plans that you could have gone to” (Twenge 2017, 53, cf. Castro and Pham 2020). And, again, it isn’t just teenage girls (or middle-aged journalists and professors) who feel this way. Nearly all of us are familiar with the fact that, now that each of us has the “option” to have a smartphone, but nearly none of us feel as though we have the option to not have one.

Simply put: many of us don’t really see a less connected life as a genuine option. And this, we think, undermines the thought that the lives we have chosen—ones that are saturated with mobile devices and the attention economy—are done with any genuine understanding of whether they are right for us.

3.3.2 Authenticity

Let us now turn to the effects of mobile devices on authenticity, exploring these effects in terms of its four facets, i.e., freedom from manipulation and coercion, consistency with the agent’s motivational states (whether they be volitional, evaluative, or long-term plans), freedom from alienation, and not being adaptive.

3.3.2.1 Manipulation

In “How Technology is Hijacking Your Mind,” design ethicist Tristan Harris delivers a laundry list of tactics product designers use to “play your psychological vulnerabilities … against you in the race to grab your attention” (Harris 2016). If there is a theme to his list, it is manipulation. Here, we’ll look at a few of the items on his list to substantiate the claim that our mobile devices are used to manipulate us.

We’ll begin with the second item on the list, “Put a Slot Machine In a Billion Pockets.” He begins this discussion—much like Eyal does his—with an estimation of how often we check our phones. His count, like Eyal’s, is 150. He adds: “Why do we do this? Are we making 150 conscious choices?” (Harris 2016). The answer he gives to the first question is variable rewards, and to the second, no. On his telling, variable rewards turn your phone into a slot machine:

• When we pull our phone out of our pocket, we’re playing a slot machine to see what notifications we got.

• When we pull to refresh our email, we’re playing a slot machine to see what new email we got.

• When we swipe down our finger to scroll the Instagram feed, we’re playing a slot machine to see what photo comes next (Harris 2016).

This is a comparison that Eyal himself is happy to make: “Variable rewards are prevalent in many … habit-forming products,” such as “slot machines and lotteries” (Eyal 2014, 9).

Turning back to manipulation, the idea here isn’t to moralize against playing slot machines. Instead, it’s to think about what a slot machine can do to a player, i.e., to get her to engage in “game play” more than is good for her. As Harris, citing NYU professor Natasha Dow Schull, warns, the variable reward aspect of slot machines is highly effective in achieving this effect (Harris 2016). In a 2013 presentation, Harris says, “These are attention casinos because the house always wins” (Harris 2013).

And, as Eyal’s playbook demonstrates, users can be made to play these slot machine-like games without being aware of the larger context (i.e., the Hooked cycle) that gets them to use the device automatically, that is “with little or no conscious thought” (Eyal 2014, 1). Here, we have powerful psychological vulnerabilities being exploited to get us to do things in ways that work around our rational agency; that is, we have a prime example of manipulation.

This form of manipulation isn’t outright deceptive, but other forms that Harris cites are. Consider list item five, “Social Reciprocity.” As we are all well aware, we have a natural inclination to reciprocate others’ gestures. If someone says “Hi,” you say “Hi” back to them. If someone waves to you, you’ll wave back. This is leveraged by product developers to get us on their sites. Harris offers LinkedIn as an “obvious offender” of this tactic:

When you receive an invitation from someone to connect, you imagine that person making a conscious choice to invite you, when in reality, they likely unconsciously responded to LinkedIn’s list of suggested contacts. In other words, LinkedIn turns your unconscious impulses (to “add” a person) into new social obligations that millions of people feel obligated to repay. All while they profit from the time people spend doing it. (Harris 2016)

He also mentions—in the context of a different but not dissimilar tactic (“Social Approval”)—how a similar effect can be achieved by auto-tagging users in photos:

When I get tagged by my friend Marc, I imagine him making a conscious choice to tag me. But I don’t see how a company like Facebook orchestrated his doing that in the first place … [W]hen Marc tags me, he’s actually responding to Facebook’s suggestion, not making an independent choice. But through design choices like this, Facebook controls the multiplier for how often millions of people experience their social approval on the line. (Harris 2016)

Here, Harris is making the plausible claim that we are sometimes deceived into using a social media service: we are given a false impression that we have a social obligation or have been noticed, driving us to the site to reciprocate or to bask in the warm glow of social approval. As we explained in Chap. 2, this is typical of manipulated behavior. Deceitful manipulation involves getting someone to act on the basis of enticements that she would not endorse if she had accurate information.

The final tactic that we will discuss is number seven: “Instant Interruptions vs. ‘Respectful’ Delivery” (Harris 2016). Our earlier discussion of attention should leave it as no surprise that messages that interrupt us immediately are more likely to get us to respond. As Harris notes: “Facebook Messenger (or WhatsApp, WeChat or SnapChat for that matter) would prefer to design their messaging system to interrupt recipients immediately … instead of helping users respect each other’s attention” (Harris 2016). As he further notes, this effect is exaggerated when messages have read receipts, that is, when the sender is notified in real time when the recipient has read the message. Leveraging the sort of psychology at play in “Social Reciprocity,” this heightens the sense of urgency and prompts a response. These forces are, again, working at cross purposes with our best judgments: many of us have accidentally left ourselves logged on and seen a message that we then feel compelled to respond to, knowing all the while that had the message been sent to us offline we would have felt no urgency to respond.

In sum, our mobile devices—often because they reward and monitor us in real time—can, and often are, used to circumvent our rational processes, getting us to jump onto sites when we in fact know that this is not what we should or even want to be doing. More often than not, when our behavior is the result of manipulation, we find ourselves disapproving of our actions or desires. We act in a way that is inconsistent with our higher-order evaluations; we feel alienated from our desires and actions. We will explore these ideas in the following sections.

3.3.2.2 Incoherent Motivational States

In the previous chapter, we explained how the coherence of motivational states is a key component of autonomy. In addition to decision-making capacity, autonomy requires authenticity. Not only must you have the ability to set and pursue ends, the ends must be, in some sense, your own. As we saw, many philosophers defend “coherence” models of authenticity. They suggest that autonomous actions and desires are ones that the agent reflectively endorses. Indeed, this is the central focus of many of the models of autonomy that have been developed in the literature. It will be helpful to recall the three accounts that were discussed in the last chapter.

For Frankfurt and Dworkin, autonomy requires consistency between first-order desires and higher-order reflection. Not only do you want to exercise, you want to have that desire. You have a higher-order desire to have the first-order desire. You want that first-order desire to motivate your action. Gary Watson’s account differs insofar as it relies on evaluations rather than desires. For Watson, the question is not whether or not you want those desires; instead, he thinks you should ask whether or not you approve of your desires. Are your desires consistent with what you think is most worthwhile? Finally, Michael Bratman’s view tells you to ask whether or not your action is consistent with your long-term plans.

With these models in hand, we are better equipped to discuss the effect of smartphones on authenticity. Perhaps the most natural starting place would be to ask yourself the following questions: First, take a few seconds to find out how much time you spend on your phone (instructions provided below).³² Are you content with your usage data? Do you approve of your first-order desire to unlock your phone when you engage in higher-order reflection?³³ Is your usage consistent with your evaluative judgments about what is most worthwhile?

You might be one of the lucky few who can say yes to those questions. But polling data show that the majority of Americans do not feel good about how much time they spend on their phones. A 2022 Gallup poll of 30,000 adults in the US found that 58% of them believe that they spend too much time using their smartphone.³⁴ This number has risen considerably since the last poll in 2015 which found that only 39% of users held this belief. In 2022, twice as many people reported that their smartphone has made their lives worse overall (compared with 2015 results). A majority of users still believe that their smartphone has improved their lives, but this might have something to do with the fact that half of the respondents said that they cannot imagine life without their smartphone (a topic we discussed above in terms of our limited options).³⁵

We have come to depend on our phones in so many ways that it should come as no surprise that they seem indispensable. But, as we explain in the next chapter, this is consistent with our view. We are not in denial about the many benefits that smartphones have to offer (calendars for keeping appointments, GPS to navigate and get traffic updates, etc.). But for many of us, even if we recognize the ways that our phones help us set and pursue our ends, we use them in ways that weaken our capacities as well. What is more, the behavior often seems compulsive. We find ourselves using our smartphones whether we want to or not. This gives us a good reason to believe that our actions and desires are inconsistent with our higher-order reflection.

Once again, Esther’s anecdote is useful. She was unable to achieve the end she set for herself (reading a book) due to her compulsive smartphone use. From what we learned above about the effect smartphones have on our capacities, we should realize now that there is more going on here than the mere opportunity cost of looking at her phone instead of the book. As Eyal explained, when we get hooked on the dopamine surge we get from our phones, we weaken our frontal lobe’s executive function, and this makes us worse at sticking to a task like reading.³⁶ This is only one of the many literacy-related cognitive capacities that is weakened by our phones.³⁷ The more time we spend engaging with the attention economy, the more our brains prefer shallow content that is less cognitively demanding.

These changes are so drastic that they can even be observed through neuroimaging. In their meta-analysis on this topic, Loh and Kanai (2014) write, “In interrupting the development of deep reading skills, this shift toward shallow information processing may affect brain circuitry necessary for these skills” (516). And in the case of what they call “internet-addicted (IA) individuals” the conclusion is even more emphatic:

Finally, the rewarding Internet environment also has resulted in the increased prevalence of Internet-related addictive behaviors. IA individuals were worse at inhibiting their responses especially in the face of Internet-related cues and were also highly driven by immediate rewards even in the face of potential losses and uncertainty. These cognitive deficits were further associated with alterations in brain networks involved in self-control and reward-processing. (Ibid.)

Esther’s desire to look at Instagram while she was trying to read a book should come as no surprise. Her brain has adapted (as have ours). She has come to prefer the dopamine-driven content of the attention economy. But this is not a simple case of shifting preferences. This is very different from a change in one’s palate.³⁸ By spending too much time with our phones and the attention economy, we have made ourselves less capable of acting on desires that we still have. We become less capable of setting and pursuing our own ends.

According to the Frankfurt/Dworkin model, Povitsky’s first-order desire to check Instagram while reading is inconsistent with her higher-order reflection. She has a second-order desire to act on her first-order desire to read (or, alternatively, she has a second-order desire to be free of the desire to check Instagram while reading). On Watson’s characterization, what is distinctive about compulsive behavior is that “the desires and emotions in question are more or less radically independent of the evaluational systems of these agents” (1975, 220). Povitsky’s smartphone use is inconsistent with her evaluative judgments about what she ought to be doing, and thus the behavior is compulsive. Finally, her action demonstrates an autonomy deficit on Bratman’s model as well. Given what she says, Povitsky, like many of us, would like to read many books over the course of her life and to develop the ability to sit and enjoy reading for long stretches. The action of looking at her phone compulsively is not consistent with her long-term plans.

Given all this information about the seemingly compulsive use of smartphones, it may seem surprising that we have, to some extent, avoided the language of “addiction.” In part, we have done this because the topic of addiction remains controversial within the field of psychology. The standard diagnostic tool for mental health professionals is the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (the DSM-5). The language around addiction has shifted considerably over time. The prime example of addiction has long been “substance use disorders” like alcoholism. Substance use disorders were added to the third edition of the DSM in 1980. Before that, alcoholism had fallen under the umbrella of personality disorders. For decades, the only “behavioral addiction” that was included in the DSM was “pathological gambling.”³⁹ In 2013, “Internet Gaming Disorder” was added as an appendix to the DSM-5. But no other behavioral “addictions” were included.

Going against the grain, some psychologists have argued that “internet addiction” should be included as well. They point out how similar internet addiction is to both gambling and substance use disorders:

Control processes are particularly reduced when individuals with Internet addiction are confronted with Internet-related cues representing their first choice use. For example, processing Internet-related cues interferes with working memory performance and decision making. Consistent with this, results from functional neuroimaging and other neuropsychological studies demonstrate that cue-reactivity, craving, and decision making are important concepts for understanding Internet addiction. The findings on reductions in executive control are consistent with other behavioral addictions, such as pathological gambling. They also emphasize the classification of the phenomenon as an addiction, because there are also several similarities with findings in substance dependency… All the findings and clinical implications discussed here have several similarities with other forms of addictive behaviors. They are consistent with neurobiological and psychological models of addiction (Robinson and Berridge 2001; Everitt and Robbins 2005) and with neuropsychological and neuroimaging findings in substance dependency and other forms of behavioral addictions. (Grant et al. 1996; van Holst et al. 2010; Brand et al. 2014, 1, 10)

We do not have a stance on whether or not internet addiction (or smartphone addiction) should be included in the DSM. Obviously, decisions of that kind should be left to the mental health professionals who standardize these terms in order to improve diagnostic and treatment outcomes. This is well outside our domain as moral philosophers.

But we are interested in the psychological literature on addiction because it offers some helpful resources when it comes to separating problematic use from that which is not. After all, many people are capable of gambling or having an occasional drink without developing a clinical disorder. When assessing whether or not someone’s gambling constitutes a disorder, here are two of the questions from the diagnostic screen⁴⁰:

• (1_d) Have you tried and not succeeded in stopping, cutting down, or controlling your gambling three or more times in your life?

• (2_d) Has your gambling ever caused serious or repeated problems in your relationships with any of your family members or friends? Or, has your gambling ever caused you problems at work or your studies?

As for substance use disorders, the DSM-5 specifies 11 criteria. Here are five that are relevant to smartphone use:

• (1_e) Consuming more than intended.

• (2_e) Persistent desire to cut down or regulate use.

• (3_e) Experiencing craving, a pressing desire to use.

• (4_e) Use impairs ability to fulfill major obligations at work, school, or home.

• (5_e) Recurrent use in physically unsafe environments.

In both cases, the parallels to problematic smartphone use are clear. First, there are questions about whether or not we have tried to cut down but have struggled to do so ((1_d) and (2_e)). For many of us, this is a familiar experience with our phones.⁴¹ Second, there are questions about whether or not the thing in question interferes with work or school ((2_d) and (4_e)). As we explained above, there is evidence to suggest that smartphones do have a negative effect on academic performance. Third, there are questions about craving (3_e) and about spending more time on these things than we intended (1_e). Smartphone use ticks these boxes as well. Finally, when it comes to using phones in “physically unsafe environments” (5_e), we need look no further than our roads. They are littered with distracted drivers who look down at their phones in spite of the evidence showing how dangerous this is.⁴²

So although we do not wish to take a stand on the clinical and diagnostic issues, we do believe that these tools demonstrate something interesting about our smartphone use. They help us see some of the ways that our devices are at odds with ends we have set for ourselves (e.g., reading books, being able to focus, succeeding in school or at work, not being anxious or depressed, driving safely, etc.). As the polling data show, some of us are already prepared to admit that we have a problem and that we would like to cut back. We would like to have more control over the amount of time we spend on our phones. This shows that our motivational states are not coherent. Much like the unwilling addict that Frankfurt described, we find ourselves doing something that we do not really want to do.

3.3.2.3 Alienated Desires

“The Making of a YouTube Radical” tells the story of Caleb Cain, who described himself as “a liberal college dropout” who was “sucked into a vortex of far-right politics on YouTube.” Commenting on Cain’s transition, a friend of his said, “I was just, like: ‘Wow, what happened? How did you get this way?’” Cain himself, in retrospect says, “I was brainwashed” (Roose 2019).

Assume for a moment that Cain was motivationally coherent after his transformation; that is, when he was—as he used to refer to himself—a “tradcon” (i.e., a traditional conservative). This certainly seems possible. He recounts being “committed to old-fashioned gender norms,” dating an evangelical Christian, and fighting with his liberal friends (Ibid.). Contrast this with Povitsky or Klein who, while behaving in certain ways—i.e., not reading in the ways they would like to—would be loath to defend those behaviors to their friends. It certainly seems that Cain was more motivationally coherent with respect to his views than they are with respect to their reading habits.

Given that Cain was (at least possibly) motivationally coherent, does this mean that his opinions were held authentically? Not at all. As we see when he refers to himself as “brainwashed,” he repudiates the beliefs upon reflection. In such cases, it is especially useful to use Christman’s historical test: evaluate whether or not we would have resisted the formation of a desire, belief, or intention if we had reflected on the process that formed it. If the desire, belief, or intention fails the test, then it is one that we are alienated from and, thus, hold inauthentically.

Cain was arguably led to his views via some such process, such as technological seduction.⁴³ As Alfano et al. (2018)—who coined the term “technological seduction”—describe it, seduction occurs in a four-step process:

• Signal. The seducer signals to the seducee, “I know what you are thinking” (Alfano et al. 2018, 300)

• Affirm. The seducee affirms, “Yes, you do know what I’m thinking” (Ibid., 301)

• Suggest. The seducer suggests, “So, let’s do …” (Ibid.)

• Agree. The seducee agrees.

Alfano et al. (2018) go on to show that this process can be carried out in technologically sophisticated ways. Relevant here is their discussion of bottom-up technological seduction, which leverages user data to propose suggestions as to what a user is thinking and what they should do. For instance:

• Signal. “[P]redictive analytics will suggest, based on a user’s profile and the initial text string they enter, which query they might want to run. For instance, if you type ‘why are women’ into Google’s search bar, you are likely to see suggested queries such as ‘why are women colder than men’, ‘why are women protesting’, and ‘why are women so mean’” (308).

• Affirm. The user affirms, say, by “selecting why are women colder than men.”

• Suggest. “[O]ne of Google’s top suggestions is a post titled ‘Why are Women Always Cold and Men Always Hot’” (309).

• Agree. The user then agrees by following the link.

This pattern need not be nefarious. However, as Tristan Harris, quoted in “The Making of a YouTube Radical,” says:

There’s a spectrum on YouTube between the calm section—the Walter Cronkite, Carl Sagan part—and Crazytown, where the extreme stuff is … If I’m YouTube and I want you to watch more, I’m always going to steer you toward Crazytown. (Roose 2019)

And this is because, as Guillaume Chaslot—former YouTube engineer, quoted in the same article—states, “it leads to more ads” (Ibid.). Alfano et al. (2021) have been able to empirically confirm certain bits of this hypothesis, showing that “there is a robust pathway [on YouTube] from some seemingly anodyne topics [such as fitness and natural foods] to outright conspiracy theories” (Alfano et al. 2021, 853). As they note, this could be “an effective way to transform people’s preferences” and behaviors (Alfano et al. 2021, 838).

Turning back to Cain, one might suspect that he would not have been as receptive to the ideas that ended up seducing him if he had known more about the pathway that he was following. He was in a vulnerable state, given that he was “[b]roke and depressed.” He slid quickly from self-help content to conspiracy theories to anti-feminist content to “explicitly racist videos.” This wasn’t a “natural” path but one that was algorithmically curated to keep him on the site. And the portability of his phone enabled him to expose himself to increasingly radical content, day and night:

That year, Mr. Cain’s YouTube consumption had skyrocketed. He got a job packing boxes at a furniture warehouse, where he would listen to podcasts and watch videos by his favorite YouTube creators all day. He fell asleep to YouTube videos at night, his phone propped up on a pillow. In all, he watched nearly 4,000 YouTube videos in 2016, more than double the number he had watched the previous year. (Roose 2019)

Now, to be sure, we cannot diagnose Cain; we do not know for certain whether he was self-radicalized in the way described above or whether he would have resisted the ideas he accepted were he to better understand the process that served them up to him. But we do think that it is entirely plausible that something like self-radicalization through technological seduction has influenced a great number of users, and some of these users would feel alienated from the views were they to know the full story of how they adopted them. Further, based on Cain’s telling of his own story, it is plausible that he is one of them.⁴⁴

Even if this weren’t the case, it seems independently plausible that mobile devices play a role in instilling habits, rituals, beliefs, and preferences in us that we would feel alienated from were we to reflect on where they come from. Consider other processes that we have discussed throughout this chapter, some of which are deployed with the explicit aim of instilling “habits” in us: the Hooked cycle, social reciprocity, instant interruption, and so on. These tactics involve getting us to do things, which might, like we think technological seduction can, get us to prefer or believe things—and we very well might feel alienated from these beliefs and desires were we to know how we came to have them in the first place.

3.3.2.4 Adaptive Preferences

Earlier, we endorsed an account that holds that a preference is adaptive if it lacks one of the following four features⁴⁵:

• (1_b) the agent reflectively endorsed the preference “at some point in its formation,”

• (2_b) her reflection took place “in the presence of recognized alternatives,”

• (3_b) “at least some of these alternatives were valuable ones,” and

• (4_b) “some of these valuable alternative options were live ones (that is, they were ones that X could reasonably see herself exercising, given her current values and ambitions” (Terlazzo 2016, 215).

Here, we will show that at least some of the attitudes we have toward our mobile devices bear the hallmarks of adaptive preferences, and we will do this by exploring the phenomenon of nomophobia (i.e., fear of being away from one’s mobile phone).

Let us begin by stepping back and talking about our attachment to our phones. A recent Gallup poll found that over 90% of Americans keep their phones near them “almost all the time during waking hours” and well over 80% “Keep it near at night when sleeping” (Saad 2022). A recent systematic review of the psychological literature on the prevalence of nomophobia found that over half of the studies reviewed (27 out of 53) determined that the rate of nomophobia in the groups they considered was 100% (Jahrami et al. 2022). And nearly all of the studies (45 out of 53) put the rate at over 90% (Ibid.). Putting this all together: it seems to be a safe assumption that many people prefer to have their phones on them.

Might this preference be adaptive? There is a good reason to think that it is.

We can begin by inspecting reasons for thinking that condition (1_b), reflective endorsement, is not met. We began this chapter by talking about some of the motivations that developers have for keeping us on our devices as much as they can and throughout it have disclosed some of the manipulative tactics they use to support the habit of keeping our phones close by. Relevant here is one of Harris’s “hijacks” that we have not discussed, Hijack #3: Fear of Missing Something Important. This hijack, which goes hand-in-hand with others—such as instant interruption, putting a slot machine in your pocket, social approval, and social reciprocity, discussed above—works by cultivating the thought that there is a chance that if you are away from an app or device too long, you might miss something important such as “messages, friendships, or potential sexual opportunities” (Harris 2016). As Odell noted above: “Media companies trying to keep up with each other create a kind of ‘arms race’ of urgency that abuses our attention and leaves us no time to think” (Odell 2019, 59).

It’s no wonder that in such an environment, we develop habits and preferences that involve keeping our devices around us at all times. Yet, as we develop these habits and preferences, we don’t often stop to think whether we endorse them. This, we think, is often by design. Many of us aren’t even aware of the Hooked cycles that are shaping our preferences as we go about our day, responding to seemingly urgent messages or checking up on our accounts to see if there is anything that we have missed. Further, even if we do, it’s not clear that we would be able to do much to stop them: being aware that cigarettes contain nicotine does not make it any easier to quit smoking once you have become addicted.

Let us now turn to (2_b), endorsement in the presence of recognized alternatives, (3_b), endorsement in the presence of valuable alternatives, and (4_b), the alternatives are seen as live. Begin with (2_b). We quoted some high school students who did not see getting a smartphone as a choice. Many adults feel the same way too. If you’re going to partake in the gig economy, you must have a smartphone. If you want to work in a variety of fields, such as academia or journalism, you might see having a Twitter account as non-optional. Further, if you are able to see not having these as options, it might be hard to see them as valuable. So much happens online, and you might be left out if you aren’t part of the conversations happening there.

Quite often, the response to those who are smartphone free or don’t have social media accounts is something along the lines of “must be nice.” This sort of resignation seems to express the attitude that such an option isn’t really on the table for me. After all, I’m not independently wealthy, or stable enough in my career (or whatever it is) for the option to be one that would, at present, add value to my life overall.⁴⁶

As a result, many of us do not see these as live options. It often feels like we don’t have a real choice. If you believe that giving up your smartphone would mean an end to your social life, no one would blame you for ruling this out as a live option. So it is no surprise that people want to have their phones on them at all times. But once we understand how this preference was formed in the absence of real alternatives, it is hard to see it as anything other than an adaptive preference. Like those who experience Stockholm Syndrome, we have come to love the devices that are holding us hostage.

3.4 Conclusion

In this chapter, we explored some of the empirical research about the negative effects that smartphones have on us. There is an abundance of evidence in favor of the conclusion that smartphones and the attention economy are undermining our autonomy. Spending too much time on your smartphone weakens your working memory and your executive function. It makes it harder to focus and sustain attention on cognitively demanding tasks. We are frequently interrupted by our phones (prone to both internal and external triggers), and we make more mistakes because of those interruptions. This makes it harder for us to set and pursue our own ends.

Our unhealthy relationship with the attention economy is making us more anxious, depressed, and sleep deprived. It worsens our self-esteem as we constantly compare ourselves to unrealistic representations of how our lives should be. Our actions and desires are being shaped by others. We may want to be the authors of our own life stories, but we often forfeit some of that self-direction as we fiendishly unlock our screens looking for the next hit of dopamine.

We also presented reasons to believe that our problematic relationship with technology was no accident. As those like Parker, Harris, and Eyal explain, tech companies have harnessed the psychology of addiction to engineer these products that we use compulsively. It is in their financial interest for us to maximize the time we spend looking at screens. Some of the successful products from Silicon Valley (like Instagram) came from students of the Stanford Behavior Design Lab, previously known as the Persuasive Technology Lab.⁴⁷ Out of charity, we may be inclined to believe that the lab’s founder, B.J. Fogg, really does have good intentions. He told an interviewer: “What I always wanted to do was un-enslave people from technology” (Leslie 2016).⁴⁸ But he is aware of how his students have used his techniques: “I look at some of my former students and I wonder if they’re really trying to make the world better, or just make money” (Ibid.). For the vast majority of products that shape our lives, the answer to this question is fairly obvious.

Once we have an adequate grasp of what smartphones are doing to our brains, our capacities, and our authenticity, we may feel moved to do something about this situation. But what exactly should we do and what kinds of reasons do we have? We turn our attention to those issues in the next few chapters where we will argue that we have a moral duty to restructure our relationship with mobile devices and the attention economy. Given what we learned in this chapter, however, we should recognize that it is likely to be an uphill battle. Unplugging is not merely a question of willpower.

As Tristan Harris points out, “The ‘I don’t have enough willpower’ conversation misses the fact that there are 1000 people on the other side of the screen whose job is to break down the self-regulation that you have” (Singer 2015). But if Kant is right about our moral duties to foster and cultivate autonomy, then we are morally obligated to resist. If we let developers set ends for us and we use products in ways that make us less capable of setting and pursuing our own ends, then Kant would argue that we are failing morally. He believes that our actions should spring from our own autonomy, rather than be driven heteronomously by “foreign impulses” (G 4:444).