‘The Social Dilemma’ Dilemma

The first thing you need to know about “The Social Dilemma”, Jeff Orlowski’s new documentary-esque movie on Netflix, is that it is inaccurate and fails to meet expectations in a fuller context, but it gets close enough that it might make some Facebook and YouTube users think twice about what is in their timeline. The other thing you need to know is that the film’s gross oversimplifications have made it an easy and reasonable target for criticism.

Mike Masnick of Techdirt accurately calls out a key omission:

Also, I should note that nowhere do they mention that Netflix, the company which funded the documentary, is also arguably the first big internet company to spend time, money, and resources on trying to perfect the “recommendation algorithm” that is at the heart of the film’s argument that these internet companies are evil. I guess some folks no longer remember, but a decade ago, Netflix even held a huge $1 million prize contest asking anyone to try to build a better recommendation algorithm.

But Masnick also links approvingly to what he calls a “dismantling” of the film by former Facebook employee Antonio García Martínez — and it’s just full of nonsense. After some mockery over the way Tristan Harris, one of the film’s interviewees, says his name is pronounced, Martínez decides he has equal parts “self-righteousness [and] narcissistic self-importance”, which is pretty rich from a guy who claims that he was the force protecting Facebook advertising from “general big-company mediocrity”. Then:

One of the first reprieves from The Tristan Harris Show is none other than Shoshana Zuboff, author of that much-ballyhooed cinder block of a book, The Age of Surveillance Capitalism.

[…]

Less diplomatically, everything Zuboff says is a nonsensical non sequitur.

“This is a world of certainty.”

Then why am I, crusty ad tech veteran, building probabilistic models all day?

What Zuboff actually said is that companies like Facebook and Google “sell certainty” to advertisers. Because they collect so much data, they promise a high degree of accuracy in ad targeting, and that helps justify their insatiable appetite for a more complete data collection picture. That is, near as makes no difference, selling certainty.

Whether they actually deliver certainty is another matter altogether. A study last year found that behavioural ad targeting caused scant improvement in publisher revenues compared to more loosely-targeted ads. That certainly suggests that attempting to pinpoint specific users’ desires for marketing purposes does not have the same meteoric impact that Facebook and Google pitch — at least in the category of many types of ads on publisher websites.

But it would be irresponsible to so quickly disclaim the possibility that the data economy and feedback loops that undergird Facebook and Google’s enterprises has little to no effect on society, just as it is ridiculous for the film to frame it as a singular driving force for increased polarization. The film cites Facebook as a catalyst for genocide in Myanmar through conspiracy theories spread by users and military leaders on the platform, but it barely mentions brutal attacks in India facilitated by viral misinformation on the chat app WhatsApp. WhatsApp may be owned by Facebook, but it behaves nothing like its parent company’s flagship product. Yet, it is possible that both products allow misinformation to spread rapidly for different but similarly damaging reasons.

Martínez mentions much of this in his newsletter, but far too easily writes off the film’s thesis. It is a polemic and, consequently, presents few counterarguments, but dismissing it out of hand means Martínez’s piece reads less like a “dismantling” — Masnick’s word — and more like an absurd justification for technologists using products they created to build the solution to problems they couldn’t possibly have anything to do with.

Adi Robertson of the Verge wrote perhaps the best critique of the film about a week before it came out on Netflix:

Big social platforms can play a key role in amplifying hate and conspiracies. The QAnon movement, for instance, began on 4chan and 8chan but spread through Facebook’s group recommendations and similar algorithmic promotional tools. In a world without gigantic, centralized social giants, its power might have been far more limited.

But small, toxic online spaces have already proven adept at hijacking a simple backlash against Big Tech by presenting themselves as alternatives. The Social Dilemma, by contrast, doesn’t offer a terribly convincing portrayal of life without its fictional Facebook clone. It assumes that if you weren’t scrolling through your feed, you’d be drawn to spontaneous tête-à-têtes or reading Zuboff’s 700-page sociopolitical tome The Age of Surveillance Capitalism… instead of tabloids, telephones, movies, television, shopping centers, pornography, talk radio, internet forums, or any of the million other activities that were previously blamed for killing good books and conversation.

The sad reality is that this is a mediocre movie about a difficult topic. It uses lazy filmmaking techniques and a knockoff of Trent Reznor and Atticus Ross’ score to “The Social Network” to frame complex problems in a far too simple light. It shows tech employees as they seek atonement, but it does not dive deeper because doing so would be a longer and dull movie. Its corny vignettes of an ordinary family torn apart by their smartphones are predictable and not helpful.

Despite all of this, it is the closest a mainstream movie has come to stumbling across a key reason online social networking platforms may be enabling our worst tendencies: we create those circumstances ourselves. Many of us habitually spend a lot of time on social media, much of which is fine if not particularly satiating entertainment. We share things we’re not certain are true because we lack expertise and cannot fact check everything that comes down the firehose. We stimulate our anger by hate-watching some cable news or talk radio asshole. All of these things are juiced by engagement predictions and financed by unethical advertising, and it is hard to argue that both of those things are directly responsible for our indulgence. But they may well be playing a part in making our worst tendencies worse. It’s hard to say for sure — that is what the film gets so wrong — but our own experience with our digital habits should tell us that we cannot assert an imperceptible impact.

Even though I think Robertson’s review is the best of the bunch, I do not think the conclusion she reaches in the paragraphs I quoted is fair. I do not think the overall expectation is that reducing targeted social media feeds from our lives will free up time to write the next great novel or begin studying Mesopotamian architecture. But perhaps a combination of stricter controls on personalization and behavioural targeting combined with antitrust regulation can reduce the outsized impact of a few Silicon Valley giants. We cannot disclaim personal responsibility for our social media diet. We also should not ignore the possibility it is being abused. We might not read more books or have more face-to-face conversations with a more regulated industry, but we might have a better chance of reading better news and having better online conversations.