It’s Time to Rethink the Assumptions That Guide the Design of Tech Platforms

The internet is getting worse. From search to news to music, it’s harder and harder to find the things you need, do the things you want, or find the information you’re looking for. Reasons for the internet’s degradation abound—the avarice and unassailable scale of tech oligopolies, misapplications of AI, the failure of regulation, and of course, the profit motive that has infected nearly every aspect of the web. There is truth in each of these diagnoses, but we believe that the current state of digital platforms is not only a consequence of market forces or technological limitations, but also of the philosophical assumptions that guide the design of these platforms.

As economists, we argue that the philosophical assumptions that guide welfare economics have also guided the design of digital spaces: specifically by defining individual well-being by the theory of revealed preferences. The fundamental idea that a person maximizes their welfare by making choices that reflect their underlying preferences remains central to microeconomic theory.

We see a similar utilitarianism in the technology industry’s approach to digital design: from streaming platforms to news sites to social media. It’s well-known by now that technology companies design platforms to optimize defined “engagement metrics”—quantitative measurements of actual user behavior, like daily active users on Facebook, time spent viewing videos on YouTube, or swipes on Tinder.

What remains unstated is the implicit logic of revealed preference theory present in the use of these metrics: time spent viewing videos on YouTube is time a user wants to spend on viewing videos on YouTube, and time a user wants to spend on viewing videos on YouTube is equated with what’s best for her—and so YouTube optimizes for time spent watching videos. In other words, it is assumed that people’s choices online reflect their actual preferences, and their preferences define what’s “good” or worth valuing in these digital spaces. It’s a convenient logic for tech companies, where what’s “best” for the user is what’s best for business.

How this philosophy shows up in your feed

Without this perspective, it can be difficult to get more specific about why the internet feels like it’s getting worse. Understanding the current state of technology through this philosophical lens allows us to better diagnose the problem—after all, the problem of what constitutes individual well-being is not specific to the internet; it’s something that philosophers have struggled with for centuries.

Take, as an example, a person scrolling his feed on Twitter/X. In this moment, this person acts on his immediate preference to consume more media, whether it’s news, updates from friends, funny videos, or the like. For tech companies, this is the only space of evaluation that matters: what kind of content he consumes, and for how long. But by drawing on a philosophical critique of the definition of well-being as preference satisfaction, we can articulate the various issues that emerge in this example.

• First, though his immediate preference is to scroll, this person might not actually include “scrolling social media” in his definition of a good life. As utilitarianism in economics is notorious for doing, this dominant philosophical approach to designing digital platforms neglects the origins and nature of one’s preferences.
• Second, his immediate revealed preference may stem from false beliefs or anti-social tendencies. This may eventually lead to self-destructive outcomes, such as belief in false claims about medicine that negatively impact his health. It may have even more destructive outcomes, such as in the various cases of young people radicalized online into committing violence, from school shooters to terrorist recruitment.
• Third, problems of confirmation bias and the tendency to seek opinions like his could cause him to gravitate towards “echo chambers” or to adopt various bigotries that lead to negative outcomes not just for himself but for society at large.
• And finally, the user’s immediate revealed preference is not nearly as self-selected as it seems; in fact, technology companies explicitly work to make platforms as engaging as possible, thereby actively forming users’ preferences online.

The popular idea of “enshittification” correctly laments the tendency of tech platforms like Facebook, TikTok, and Google Search, once an audience is sufficiently enclosed, to sacrifice content personalized to what you want to see for the content that the platform wants you to see. But we question whether even designing for what you want to see, in the sense of revealed preferences, is a suitable goal. We think this perspective may offer a better explanation of what’s happening on streaming platforms and social media sites in the era of “AI slop.”

For instance, Netflix and Spotify optimize for “casual viewing” and “lean-back listening,” respectively, not only because these add to their bottom lines, but also because they serve what their engagement metrics tell them users prefer. This leads to increasingly poor content and experiences, but is purportedly based on what users want, derived from their observed behavior.

The capability approach as an alternative for technologists and policymakers

If the use of engagement metrics in digital design has been criticized more generally, a popular alternative in the tech industry is the use of qualitative measurements, such as user satisfaction, perceived ease-of-use, and net promoter score. Yet we view this approach as merely a variation of the utilitarianism we have critiqued so far—in which utility is conceived of as a mental state, such as happiness or pleasure, rather than revealed preferences—and therefore falls into many of the same traps.

For one, self-reported mental states reflect individual variation in natural disposition and therefore obscure objective disadvantage: someone who is naturally agreeable or who has adapted to her given situation (for example, low vision or limited access to the internet) may report the same levels of online well-being as someone without these disadvantages in objective circumstances. Additionally, these qualitative measurements, like their quantitative counterparts, do not necessarily involve a process of critical self-reflection, especially given the constant requests for this kind of feedback today.

If this utilitarian approach to digital design is found wanting, we argue that the Nobel-prize winning economist Amartya Sen’s idea of capabilities is a promising alternative. In his economic and philosophical work, Sen criticizes utilitarianism both for the position that what is right is whatever maximizes welfare (however welfare is defined), and the specific version of it that defines welfare as preference satisfaction. In contrast, Sen emphasizes that an individual’s observable behavior represents just a subset of her capabilities (what she has the potential to do or be). In this way, we can expand the space of evaluation from utility to “the substantive freedoms—the capabilities—to choose a life one has reason to value.”

This approach’s advantage has to do with how it incorporates agency and freedom: it asks us to not just defer to people’s revealed or stated preferences, but rather to distinguish between people’s preferences and their scrutinized values, to consider not just what people do but what they can do, and to require that people critically evaluate and deliberate about which of these capabilities are of value in these spaces. As the philosopher Mark Sagoff once put it, “democracy does not take preferences as they come but alters them; for example, it subjects them to public scrutiny and debate. At its best, democracy works not by aggregation but by deliberation.”

To return to the example of the person scrolling his social media feed, the capability approach expands the space of evaluation beyond his revealed preference to scroll. Because Sen incorporates concern for agency and freedom, the prescriptive and singular experience of endlessly scrolling his social media feed becomes an unfreedom in itself: this person is unable to freely choose between alternatives in this space, as personalization algorithms continually narrow his experience to his revealed preferences. Moreover, this person is unable to freely choose even between alternative platforms, as Big Tech reinforces its monopoly over digital spaces through network effects and barriers to interoperability.

If we consider individuals’ capabilities, not just limiting ourselves to revealed preferences in digital spaces, we can acknowledge that the example of satisfying a person’s revealed preferences to scroll their social media feeds crowds out other forms of social interaction in real life—consuming more of their limited time and mental energy, to the point that it’s almost a cliché now to say that the iPhone and social media have impacted young people’s mental health and increasing loneliness in the West. Currently, it appears that technology companies only consider the idea of welfare or “flourishing” insofar as use of their products becomes an end in itself, reflected in the pursuit of perpetually improving engagement metrics, rather than a means to a life well-lived. We see the capability approach as allowing technologists and policymakers to embrace the idea that the internet should help us achieve a good life.

Conclusion

Why should we care about the philosophical underpinnings of digital design? Our argument is that the capability approach offers practitioners and regulators a better system to evaluate what companies do and should do—whether in the practice of design, in industry regulation, or in the creation of public platforms.

While our diagnosis applies to many kinds of technology, from cultural products to news sites to social media, we argue that the use of the capability approach as a positive program is best suited to digital public spaces of public interaction, conversation, learning, and deliberation like social media—namely, the public spaces of the internet. That’s because while the market economy in liberal democracies generally denies consumers the right to directly decide upon what products a company sells and how, we do maintain the right to say how we interact with each other, whether online or offline. The capability approach reminds us of the context in which we exist online—broadening our concern from “How do we live a good life online?” to “How do we live a good life?”