Terms of Disservice - Dipayan Ghosh

Скачать книгу

take our understanding of the disinformation problem exclusively from their statements, we would not suspect the industry itself as a primary culprit. But it would be foolish to ignore the possibility that this problem is a by-product of Silicon Valley’s pell-mell pursuit of mammoth profit margins by expanding domination across the territory of the digital landscape with reckless abandon—or, to rephrase Facebook’s old motto, by “moving fast and breaking things.”26 After all, we did not have this problem before the age of vast internet commercialization; we were not plagued by misinformation and foreign election interference in the ways that our political communication networks were infiltrated in 2016.

      It is on us to dissect what internet executives have thus far said and take it for what it largely is: commercial propaganda. The reality is that the economic infrastructure that defines the internet is centrally responsible for the widespread damage that has been done to the American media ecosystem.

      Of course, as Sandberg suggests, nefarious actors are working for the Kremlin and beyond and have it in for us and are eager to develop new methods for injecting opportunistic political messages into the American discourse. In fact, we may never be able to make them disappear completely; even if the U.S. government were to organize a clever set of sanctions that debilitated Russia’s intent and capacity to attack our political and information systems, other actors would quite likely spring up and aim to subvert the sovereign strength of the United States. We have already seen some evidence of information operations emerging from China.27 That this is happening should come as no surprise; such is the way of a multiethnic world replete with transnational economic competition and the resultant geopolitical uncertainty.

      But what did not exist until recently was the current mode of internet commerce—led by the likes of Google, Twitter, and Facebook—that has promoted the reemergence of digital propaganda and other classes of harmful communications. It is the creation and facilitation of the novel influence market hosted by the consumer internet firms that is responsible more than any other single factor for the prevalence of this series of public harms.

      We need to step away from the industry executives’ injections of engineered noise and distill a comprehensive policy regime—a new internet order—that can once and for all motivate a truly earnest conversation about what the United States and jurisdictions around the world should do to contain the capitalistic overreaches of the consumer internet industry.

      Unveiling the Web of Secrecy behind Digital Commerce

      To design the appropriate regulatory intervention to contain Silicon Valley’s public harms, we must start with a thorough analysis of the business model that is in play. At its heart, that business model is joltingly simple: it involves the creation of compelling internet-based apps and services that limit competition over the internet, as well as the uninhibited collection of fine-grained information on individual consumers to create a behavioral advertising profile on them, and the ongoing development and implementation of a set of algorithms, including artificial-intelligence systems, that automatically curate social content to engage consumers and target ads at them in a programmatic manner.

      My contention is that given this business model, we need aggressive, reformative policy regimes that can better ensure individual privacy, increase consumer transparency, and promote market competition in the consumer internet sector. It is only when examined through the lens of the consumer internet’s business model that the purpose of the industry’s corporate decisionmaking comes to life, enabling the type of truth-busting skepticism necessary to keep this industry and its executives honest.

      Why was the Cambridge Analytica incident—in which Facebook data pertaining to some 87 million users was compiled and illegally sold to the British political consulting firm that was contracted to Donald Trump’s presidential campaign—a critical event from a technical perspective? Was it really just like the data breaches involving Capital One or Target or Sony or the hacking of Experian customers’ e-mail addresses and financial information? Resoundingly, no.

      Experts have discussed at length the troubling capabilities that Cambridge Analytica advertised. It could, for instance, engage in psychographic analysis and categorize people along five psychological qualities—openness, conscientiousness, extraversion, agreeableness, and neuroticism—determined by their behaviors and activities as revealed through their Facebook engagement data, including “Likes.” This is not a unique capacity. IBM, for instance, has used its Watson artificial-intelligence platform to generate detailed classifications of psychographic traits and apply them to text.28 Nielsen has advertised similar capabilities. And, in fact, it has widely been suggested that Cambridge Analytica’s capabilities in driving psychographic inferences on users for client applications in the political context was limited, at best.29

      But beyond any capacity Cambridge Analytica may have had in developing psychographic conclusions about a given user, I think the most critical aspect of the incident involved the potential breach of personal identifiers, possibly including what are known as Facebook User IDs—individually identifying pin numbers that would enable whoever has the Cambridge data to create vast (or narrow) target audiences for digital advertising campaigns on both Facebook and non-Facebook platforms and to coordinate highly effective disinformation campaigns. Most advertisers on Facebook have to go through the hoops of mapping known customers to their Facebook accounts, which is a process that does not always yield many matches to Facebook profiles. However, if the Cambridge Analytica breach included user IDs, the mapping would have been done for them. In other words, instead of reaching a matching yield of just 30 percent or 50 percent, the firm’s clients could reach 100 percent of the 87 million accounts whose data the firm had obtained. Given that the vast majority of the 87 million accounts were those of American voters, the firm essentially had access to a targeting-and-tracking regime via Facebook’s advertising platform that would have enabled a degree of illicit targeted political communications the likes of which we have never seen before.

      While we might have some clues from independent researchers about what data may have been a part of the breach given analysis of data generated by Facebook’s old application programming interface (API),30 the company has provided very little in the way of public guidance to address the actual data exposed in the breach. I have yet to see an in-depth analysis of the Cambridge Analytica incident that highlights the dangers associated with the breach of personally identifying Facebook data that potentially included user IDs—despite the reality that this likely is the single-most critical facet of the Cambridge Analytica incident.

      That Facebook consistently fails to address the implication that specific users could readily have been targeted on the company’s on- and off-platform advertising services using the identifying data obtained by Cambridge Analytica is highly discouraging. It composes precisely the sort of obscurantism that deserves great public scrutiny and thorough investigation. Consider, also, Facebook’s recent decision to enable texting across its three major internet-based text-messaging services—Instagram, Messenger, and WhatsApp. Each enjoys hundreds of millions of users, which for Messenger came about in large part because Facebook some years ago forced many existing users who had Facebook accounts to download Messenger if they wanted to text over Facebook with their Facebook friends. On its surface, Facebook’s decision to enable this cross-messaging might appear to be a boon to consumers—who would reject the efficiency and ease of maintaining contacts on a single service versus switching back and forth between apps to keep in touch with friends?

      However, what I find most striking about this decision is the way it is presented to the public by the company—as a technical change that will offer great convenience to customers. Of course, this might be one of the factors that drove the company to make this change; there was every possibility that by integrating the three applications users would find their experience to be more convenient and would therefore communicate with each other to an even greater extent over the three services or save some time in their day by not having to switch between mobile applications and websites.

      But if everyone who uses internet messaging is a user of any of the aforementioned three apps and therefore can message with anyone else who is also a user of one of the three networks, what could drive the scaled adoption of a new texting service that might someday compete with Facebook’s universe of texting networks? As such, this advent would also accomplish something starkly

Скачать книгу