Wednesday, May 15, 2013

To Save Everything, Click Here

Evgeny Morozov, To Save Everything, Click Here: The Folly of Technological Solutionism: Morozov’s excellent first book, The Net Delusion, established him as a major critic of internet eschatology, particularly the utopian brand. His latest book attempts to expand on that critique, sometimes successfully and sometimes not.  For a sample mostly taken from the book, here’s Morozov on why we shouldn’t be sanguine about using technology to “improve” incarceration.   A good sampling of Morozov’s work is available online; I would recommend avoiding his debate with Farhad Manjoo in Slate, where Morozov mostly ignores legitimate hits (he attacks generalizations like “the internet” but himself has no trouble criticizing “Silicon Valley”) in favor of snide near-ad hominems.

Morozov wouldn’t be surprised that I, as an internet reviewer (“ordinary people don’t write reviews for the same reasons as professional critics; they are mostly interested in reviewing their own experience, not in making sense of a given work”), can’t improve on Kevin Driscoll’s take in the LA Review of Books, which has some very smart things to say about Morozov’s critiques of “solutionism” (social issues as problems with a fixed solution rather than approaches that have to be negotiated and compromised on) and internet-centrism. For a rather harsher view, there’s Tim Wu, who is attacked in the book and understandably annoyed; his criticisms are not unwarranted.

Morozov is a skeptic not just of whether Silicon Valley’s big promises can be carried out but, more importantly, whether they should be. Friction, lack of transparency, inefficiency, and so on are not just hinderers of efficiency, but crucial parts of human self-definition and autonomy; politics isn’t politics if it’s not messy and a bit hypocritical. Compromises and imperfections can be good, not bad; politics can’t be improved the same way that market transactions can be—if all interactions could be win-win, we wouldn’t have politics. “Try telling [an Amazon] shopper that not all of his or her desires can be satisfied because someone else has equally compelling interests and those have to be taken into account as well; the market simply doesn’t work that way.”

And the consumerist mentality that solutionists bring to political challenges leads to disappointment and disgust with “politics,” when it should lead to a rejection of solutionism: “Most public institutions should not be held to the same standards as their private counterparts because their mission is to provide goods and services that markets cannot or should not provide.” More generally, inefficiencies, hypocrisies, and the existence of crime “might be problematic in some limited sense, but they do not necessarily add up to a problem worth solving—any more than having a soccer match that lasts for ninety minutes rather than an eternity and features twenty-two people instead of everyone at the stadium is a problem to be solved.”

Relatedly, the “frictionless” solutions promoted by techno-utopians often don’t solve the problems they purport to, and they don’t solve them in particular ways reflecting existing political and social inequalities. For example, biometric identification technologies turn out to have particular trouble with certain racial groups, and fingerprint scanners have difficulty with people in certain working-class occupations--not that this is anything new, as Morozov is at pains to point out.  (An example related to my own field is the way in which visual communications technology is bound up in whiteness: Richard Dyer, "Making 'White' People White, in The Social Shaping of Technology, eds. Donald MacKenzie and Judy Wajcman (noting that, among other things, videotape quality was evaluated by how well it displayed a blank, pale orange signal called "skin" that was supposed to match white skin); Brian Winston, A Whole Technology of Dyeing: A Note on Ideology and the Apparatus of the Chromatic Moving Image, Daedalus, Vol. 114, No. 4 (Fall, 1985), pp. 105-123 (discussing how, at every stage, film development was guided by how it did at showing white skin).)

For another example of solutionism, massive online open courses, Morozov notes, as many others have, that what they offer isn’t individual contact with an expert (even a grad student) but rather something else, and they don’t exist in a political vacuum: “In promising almost immediate and much cheaper results, they can easily undermine support for more ambitious, more intellectually stimulating, but also more demanding reform projects.”

Algorithms to recommend books to you have their biases, but that’s nothing compared to the dangers of algorithms that predict crime based on current circumstances (including poverty and racism that mean that crimes are committed and detected in particular ways). And design that simply prevents the possibility of crime also often prevents the possibility of civil disobedience, an important driver of social change: “Sometimes being caught with marijuana in one’s pocket is better than being prevented from putting it there, simply because an arrest is likely to generate media attention and trigger a public debate about drug laws.” But he’s always context-sensitive, and not dismissive of all better living through technology—anti-drunk-driving technology might be a good idea even if other technological interventions aren’t.

I was reminded of Morozov’s critique of big data solutionism by this bit in Slate about how facts can change their meanings when you have more context:
Marti Barletta, a consultant in marketing to women, told me one of the reasons women have gained a reputation for caring about frivolous details is because they do so much research. By the time they arrive at dealerships, they’ve already logged countless hours online finding cars that satisfy their main criteria. Now, they’re picking through minutiae—what, precisely, makes the Nissan Maxima better than the Toyota Camry? (Could it be the number of cup holders?) These questions, Barletta says, contribute to an impression among salesmen that women care mostly about the little stuff.
This is a good example of the point that “data-driven solutions” can’t ever be entirely data-driven: you always need a theory.

Morozov also constantly emphasizes that people construct technologies. There’s nothing inevitable about the configuration of “the internet,” or “Facebook” or “Google” for that matter. Facebook could limit the number of ads it shows; Google could write different algorithms. As Tarleton Gillespie insightfully noted, Twitter’s algorithm for picking trending topics favors breadth (many groups using the same hashtag) over depth (a united group using the hashtag a lot), which is a political design choice that can be contested. When we focus on current numbers, such as monitoring how much water we consume, we may be motivated to take individual action but we don’t understand or consider the complex systems of overall water consumption, and we aren’t challenged to think of how we might get a different set of numbers—Morozov wants our technologies to confound and challenge us.

This is an important reminder, but it leads Morozov to be highly critical of activist discourses around things like SOPA/PIPA/ACTA’s “don’t break the internet” advocacy, and here (like many of his reviewers) I think he somewhat misses the mark. Very few of the leaders, and I suspect very few of the followers, of the anti-SOPA/PIPA/ACTA protests believed that http and IP addresses and the like would disappear under SOPA/PIPA/ACTA etc. Rather, they believed that key features of the internet they knew and liked would be hampered if not destroyed. The “internet” that existed, they thought, would be similar to the “Medicare” that would exist if what we now call Medicare were replaced by a voucher system. It’s fair game in such a debate to say that “Medicare” would be destroyed by such a change; it’s even fair game to say that “marriage” would be destroyed by extending it to women who want to marry women, even if that’s a dumb argument on its own merits. Yes, of course we should often question definitions, where much of the important rhetorical work is getting done, and Morozov is right to point that out—but many of the people he accuses really do know that already, and have made it pretty clear that they do.

(He’s particularly unfair, it seems to me, to Jonathan Zittrain, whose The Future of the Internet and How to Stop It comes under attack for internet-centrism and quixotic desire to keep the internet in a single state forever. E.g., Morozov says “to claim that Apple—one of Zittrain’s culprits—is bad for innovation because it’s bad for ‘the Internet’ is like claiming that ‘the Internet’ is bad for innovation because it’s bad for the telephone. Well, it might have been bad for the telephone—but when did preservation of the telephone become a lofty social goal?” Yet a significant chunk of Zittrain’s book is devoted precisely to addressing questions of when we can say a technological configuration/change in the direction of greater or lesser centralized control is a good thing. Zittrain isn’t entirely successful, I think, but by failing to acknowledge his explicit attempts to grapple with Morozov’s points, Morozov makes it seem as if Zittrain were silly or hypocritical, and himself ends up fighting a straw man. This ungenerosity isn’t unique in the book. Later, for example, Morozov suggests that those who favor market transactions trading private data for material benefits as mutually beneficial must also therefore approve torture “provided the prisoners ‘strike the right deal’ and are well compensated,” whatever that might mean.)

Morozov is at his best discussing tradeoffs and political reactions to technology: greater access to information can be manipulated by governments just like other new forms of power can be, and if—which is not yet established—there is a link between the two, he’s right that it’s not clear that the “local politics in Bahía Blanca [Argentina] [should] make sacrifices so that a fifteen-year-old in Palo Alto can remix cat videos without going to jail.” He doesn’t want that fifteen-year-old to go to jail, but he also doesn’t want arguments for that remixer’s protection to prevent any tinkering with technologies to make them “safer,” for some politically chosen definition of safer, especially since private parties and nondemocratic states are willing to tinker anyway.

Some have found his attack on extreme self-monitoring technologies and technologists, who are really outliers, to be a bit much, but I really liked his point (again, not new, but well made) that there is a deep political problem with proposing self-monitoring as the solution to the barrage of advertising and subsidies that keep us eating terrible, unhealthy food: “yes, some of us might find ingenious engineering solutions to resist insidious marketing, but in all this celebration of modern technology, shouldn’t we also do something about the marketing itself? …. [P]olitical action all but disappears; rather than reforming the system, we just tinker with ourselves and tend to our reservoirs of willpower the way Swiss bankers tend to their vaults.” But Morozov also hates “nudging” via technical or legal structures, even though that’s pretty much the opposite of the individualistic solutions he condemns, because he wants us all to think deeply, and exhaustingly, about all our politically relevant choices, which is to say basically all of them, though he talks most about energy consumption.

Probably Morozov’s most effective assault is on the concept of “openness” as an unqualified good. Because of preexisting political struggles, “open” data will be used in politically inflected ways—for example, maps that visualize crime statistics across different neighborhoods could help improve police effectiveness, but they could also devalue properties and make people living in dodgy neighborhoods to be less willing to report crimes. Openness has feedback effects.

Likewise, digitization of land records in India, in an attempt to empower the weak, may have benefited the rich and powerful by exposing which occupants lacked formal title despite being morally and even legally owners. Morozov advocates for context-specific solutions—here, accepting other methods of proving title such as old family photos or maps along with official land titles, or selectively limiting access to land records so that people with “no obvious need” to see them can’t do so. Information, he argues, should be “collected and distributed in full awareness of the social and cultural complexity of the institutional environment in which it is gathered. Sometimes preserving the social relations that enable that environment to exist … might require producing data that is only half transparent or half accessible …. [D]emocracy thrives on compromise and the art of reconciling seemingly irreconcilable interests.” And it’s hard to disagree with that last point, whatever excesses are in the rest of the book.

1 comment:

Anupam Chander said...

Rebecca, thanks for a really great review. I am skeptical even of Morozov's first book, which earned such acclaim, however. See my essay in the Cornell Law Review last year--
and a short version of it posted to Legal Workshop this week: