Editor’s note: Deepnews.ai is a technology company, though we also like to look at conversations happening around the use of algorithms to highlight information online. This is one of the occasional posts we do about discussions happening on topics that we find interesting.
By Christopher Brennan
How do you solve a problem like social media? How do you even show it?
“The Social Dilemma,” a new film from director Jeff Orlowski, has been a frequent feature on Netflix screens since it came out earlier this month, and has been among the most-watched in many countries.
More than something to watch, it has also become something to talk about, which is why this week’s blog post (after the good folks at Netflix politely declined to make Orlowski available for an interview) is on the movie, the attention, the criticism and what it all has to do with Deepnews (a lot!).
The film’s basic structure is centered on interviews with critics of major social media platforms, often former employees such as ex-Google design ethicist Tristan Harris. These are interspersed with scenes from a fictional storyline where a teenager named Ben is toyed with by social media algorithms, personified by three identical men who send notifications, recommend videos, and push him towards certain types of content until he becomes increasingly radicalized.
While many have praised the film for taking shots at everyone’s favorite companies to criticize, others such as Casey Newton (of The Verge but now departing for Substack) have knocked it for a “cartoon super villain view of the world.” He and his colleague Adi Robertson also point out that the film does not address the problems beyond recommendation algorithms, such as misinformation spread by WhatsApp or radicalization on small chat forums.
This is true, though I disagree with Newton’s claims that filmmakers took a “cop out” by blaming a lack of nuance on the fact that they were targeting a large audience. This, to me, misses the point that the film succeeds in drilling down on one aspect of the social media experience and making the idea stick. Tech journalists of course are more than ready for in-depth articles about the way that the complex problems of tech, media and society fit together (the sort of articles that the Deepnews Scoring Model often rates very highly). But this film is not particularly for them.
Its intended audience brings to mind another, less weighty “dilemma” sometimes faced when writing articles on tech for a general audience: which image to use for a story about something abstract. Oftentimes for tech or business-site stories about topics such as moves against bad actors by Facebook or Twitter, the image above the text will be a file photo of a Facebook logo, Jack Dorsey, or maybe even a dull shot of Google’s campus. However, if you spend time on local news sites from Peoria to Palm Beach where the audience is following things around Palo Alto less closely, you may be more likely to encounter a very particular type of stock image, the glowing computer screen and some hooded figure, representing a “hacker” or malicious cyber actor. These are designed to connect to an audience of people who don’t read the MIT Technology Review and who don’t even normally click on a tech-centered article because it is not the subject they are interested in.
These pictures can sometimes be ridiculous, but to me the difference between how tech sites and local news approach stories reflects the same divide between “The Social Dilemma’s” mission and its critics. Like local news, the movie, through its particular medium, is meeting people where they live. This is in their personal lives, rather than ongoing narratives about what Facebook is doing on complex questions of content moderation or platforms’ effects on the advertising economy. This is why there are no extended meditations on figures in the industry in “The Social Dilemma,” even the often targeted Mark Zuckerberg. While the speakers in the movie have inside knowledge of the decisions being made, the film operates almost entirely on two levels: the personal level that touches on the psychology of how we interact with social media, and the pie-in-the-sky level of “society” that attempts to explain how it is changing the world with its scale.
Orlowski and his guests are not bringing any particularly new information on either of these levels, but one thing I do think he does well is unite them through the theme of driving engagement, which is where we start to touch on Deepnews.ai a bit. In the movie, this theme is largely explored through the visualization of the recommendation algorithm, the three identical men in a room prodding Ben with news of his ex-girlfriend or appeals to his political alienation. While the recommendation algorithm is not the source of all problems with online communication, it is a good place to collect disparate strands of the tech universe and Orlowski’s view is more accessible than trying to make a movie about academic topics like Shoshana Zuboff’s “logic of accumulation.” While “The Social Dilemma” is not particularly nuanced, to me it does succeed in using its 90 minutes to go beyond the message of “social media = bad” to one that says “the structure of the way we do social media is heading in a bad direction.” If it were summed up a phrase, it would be Center for Humane Technology’s Aza Raskin saying that the industry is “tilting the floor” towards bad outcomes: the misinformation, addiction, etc., that make up separate issues but in many ways are tied together.
While the “The Social Dilemma” does drive home an important message in a blunt way, the criticisms of the movie I most agree with are the ones that point out its lack of a way forward. Towards the end of the film the interviewees bring up the usual suspects of suggested changes: “Regulation!” “Change the business model!” “Make it unprofitable to do bad things!” While film’s exploration of the “bad direction” is fairly focused on engagement at all costs, the solutions section sometimes feels unrelated to the problem or mushy and vague, with statements like Harris mentioning that technology applied differently could “make the world a better place.”
This, to me, is where Deepnews.ai would fit right in. The filmmakers and participants don’t give any answers as to what a “good direction” could look like, either because that is not their real project with the movie or because it is more difficult to find the equivalent of glowing keyboard hacker images for something positive. Deepnews.ai, to me, is the perfect example of an effort to take technology in a different direction. Our algorithm does not strive for engagement, and is not creating fragmented realities for our readers. Rather, it is focused on surfacing content based on its inherent qualities, the interplay of ideas and the value added, the amount of work that went into the piece. While the current model of recommendation aims at engagement at all costs, the place where Deepnews “tilts” is one of common facts, rigor and deliberation, which perhaps could be useful for reversing some of the damage to democracy that we’ve seen.
The problem is, once again, how to show this. As our CTO Girish Gupta mentioned last week, it is hard to get people to pay attention to what an algorithm itself is doing and how it is different. This is part of the idea of our upcoming tech demonstrator, that we want people to be able to look at the type of results they receive from Deepnews.ai in comparison to the type of results they would receive elsewhere. When it comes down to it, driving home a message means getting people to engage. “The Social Dilemma” has succeeded in doing that by offering up dystopian views of the world. While it is most engaging to drum up outrage over the problem, the real task now is engaging with solutions.