Category: Culture

Big Data in politics. Maybe we should break the internet.

Photo by LisaD

If it feels just a little bit like the world is careening toward the edge of a cliff with a madman at the wheel, maybe it’s because that’s what’s happening. Except the madman isn’t just some garden-variety berserker.  It’s not President Trump with his incoherent tweets and unabashed lies. In fact, according to this in-depth story by Carole Cadwalldr, writing for The Guardian, those antics are calculated theater being fed to the press as just one component of a much larger, more insidious process by which money and computing power are undermining democracy itself.

Unfortunately, it turns out that the madman at the wheel is us.  All of us. Voluntarily feeding the database via social media, teaching the system exactly how to tell us what to think. “There are two things, potentially, going on simultaneously: the manipulation of information on a mass level, and the manipulation of information at a very individual level. Both based on the latest understandings in science about how people work, and enabled by technological platforms built to bring us together,” Cadwalldr writes.

Focusing primarily on billionaire/computer-scientist Robert Mercer, his relationship to Stephen Bannon and the development of the Breitbart network, Cadwalldr details the process by which wealthy, mostly right-wing, individuals are using computing power not to understand the electorate (that would be old-school) but to manipulate the electorate into reshaping the world as these individuals believe it ought to be shaped. Cadwalldr’s is one of several articles to appear in recent weeks about the role of Big Data in global politics, and just one to mention the company Cambridge Analytica. She writes…

“On its website, Cambridge Analytica makes the astonishing boast that it has psychological profiles based on 5,000 separate pieces of data on 220 million American voters – its USP is to use this data to understand people’s deepest emotions and then target them accordingly. The system, according to Albright, amounted to a ‘propaganda machine’.”

Cadwalldr cites several scientists who voice deep concern about the capacity of the AI to learn about us through interactions as apparently innocuous as Facebook Likes.  With the input of 300 Likes, Cambridge Analytica claims that the computer can understand us “better than we understand ourselves.”  And because the computer never stops learning, never stops adapting, this provides the opportunity for political operatives to steer the electorate toward targeted conclusions about issues or candidates.

The irony, of course, is that this is all based on the illusion that digital technology provides us with more choices to get “better” information, which is why discrediting the “mainstream media” is a key component of the strategy. But lest we believe press-bashing is exclusively a Trumpian phenomenon of the moment, we should not forget that the “mainstream” has previously been dismissed by the left and libertarian-leaning, techno-utopians of Silicon Valley as well—all singing from the hymnal that the internet is the greatest tool for democracy ever invented.

Regular readers know that I am more than a little cynical about this generalization and that this is one reason I remain critical of “digital rights” groups fearful of any form of regulation in cyberspace—particularly regulation meant to protect or restore the rights of individuals. It’s not that I find fault with the premise that the goals of openness and free speech should be protected online so much as I balk at the assumption that an absolutist approach (i.e. law has no place on the internet) can only have salubrious results for democratic values.  I believe this sensibility is a holdover from Silicon Valley’s more hippie-like early days but is a vibe that no longer has any relationship to the advertising and data-mining systems the major companies have built.

But now that we’ve mostly “left the internet alone,” allowing these companies to collect and sell information about us without any kind of rules—allowing these same companies to monetize works of authorship and social interactions without restraint—all in the name of “freedom,” we are apparently teaching the machine to effectively democratize democracy out of existence. Cadwalldr quotes Jonathan Rust of the Cambridge University Psychometric Centre thus:

“The danger of not having regulation around the sort of data you can get from Facebook and elsewhere is clear. With this, a computer can actually do psychology, it can predict and potentially control human behaviour. It’s what the scientologists try to do but much more powerful. It’s how you brainwash someone. It’s incredibly dangerous.”

The manipulation skews right. For now.

Take a subject like immigration policy and the fact that many of us who live or work in diversely populated urban centers (traditionally liberal) can’t understand why Americans who live in more homogenous, rural communities (traditionally conservative) are so concerned that Muslim refugees pose a substantial threat to security despite a lack of evidence to support this fear.

It’s not because citizens are uninformed, it’s because they are purposely misinformed by a very sophisticated network of well-crafted, smartly-written articles that contain elements of truth glued together by rhetorical paranoia.  This is in fact the structure of the average Breitbart article on immigration; and these articles become the foundation of a million ways to automate the spread and repetition of an anti-immigrant message until it morphs in the minds of readers from emotional xenophobia to what is perceived as rational security policy. This is why labeling support for an EO immigration ban as “racist” sounds absurd to many and why their response will be, “It’s not racist. It’s just common sense. Look at the mountain of evidence! The mountain of evidence the MSM isn’t reporting!”  (Never mind that the mountain is a hologram.)

This phenomenon is hardly restricted to the political right, though Cadwalldr observes that the money and institutions behind this level of big-data manipulation is largely a right-leaning agenda at the moment. My own concern has always been that these manipulation tools, neatly disguised as “democratized” information, can be wielded by any entity with the resources.  If the pendulum were to swing from Breitbart to Google or to some left-leaning billionaire’s project, that still wouldn’t be democracy.

The ability to create the appearance of consensus through rapid replication, a network of “alternative news” sites, and bot-swarms, all emanating from a single source is exactly the concern that launched this blog in 2012. The illusion that hundreds of articles or millions of people all “agree” on a given topic can be conjured by a relatively small and nimble group of people with the money and computing power to do the job.  I alluded to this concern in this post in 2012, suggesting that Citizens United was child’s play compared to the capacity for manipulation of the political process on a one-to-one basis via social media.

In 2011/12, this sophisticated kind of disinformation was what I believed was happening with a bill like SOPA.  Now, it’s the same scenario on a much larger scale, influencing the governments of the world. The implication is the destruction of democracy by means of the very tools that were supposed to improve democracy. And the irony is that it’s all voluntary.  Every day, we get up and feed the beast. We teach the machine how to manipulate us.  So, is the only solution to abandon Facebook et al — to stop feeding ourselves as data to the machine?  Can we even afford to unplug given that these platforms are now almost indispensable for access to information and to substantive interaction with people?  A Columbia Journalism Review study on the effects of the Breitbart media ecosystem, offers these words of wisdom and hope:

“Rebuilding a basis on which Americans can form a shared belief about what is going on is a precondition of democracy, and the most important task confronting the press going forward. Our data strongly suggest that most Americans, including those who access news through social networks, continue to pay attention to traditional media, following professional journalistic practices, and cross-reference what they read on partisan sites with what they read on mass media sites.” 

So, maybe we don’t have to break the internet so much as break a lot of really bad habits the internet keeps trying to teach us.

Also read:  The Rise of Weaponized AI here.

How Bad is the FCC Pause on Privacy Rules?

Photo by kentoh

By now, you may have encountered a handful of headlines stating that the new, Trump-appointed FCC Chairman Ajit Pai has temporarily halted new privacy rules passed under Obama’s Chairman Tom Wheeler.  As a general takeaway, you’re also likely to see statements like “Republicans favor corporations over consumer privacy,” and as with all things under the Trump imprimatur, trust is not going to be the default position for many of us. But, then, viewing every issue through the lens of Trumpism is also another way of distorting reality, making ourselves bananas, and failing to compartmentalize issues, which is almost always necessary.

On the subject of consumer privacy and the FCC, any criticism that Wheeler’s policies favored companies like Google and Facebook over other companies should be at least considered.  As readers know, I and others criticized Wheeler’s so-called “set-top-box” proposal as an attempted hand-out to Google at the expense of television creators.  So, when this story about privacy broke, I wasn’t prepared to take every headline at face value just because Pai is a Republican in the Trump administration. He was in the FCC before the election and would likely still be there regardless of who became president.

Let me jump to the main point here:  we consumers want and deserve control over the gathering, selling, and use of our data.  We do not have this control yet. There is no comprehensive, uniform body of law that is adequately protecting privacy in cyberspace for American citizens.  And these facts are central to Pai’s criticisms of Wheeler’s rules of October 2016.

With the adoption of NetNeutrality—another topic for another day—the FCC was empowered to regulate ISPs like AT&T and Verizon in the same way it regulates telephone carriers.  As a result, it was then empowered to establish rules for these ISPs regarding the manner in which they may gather and use consumer data.  This is fine in principle, but critics, including Pai, argued that these FCC rules were not consistent with the rules set by the Federal Trade Commission, which apply to data gathering and use by companies like Google and Facebook, referred to here as “edge providers.”

Because the ISPs hope to compete in online advertising with the “edge providers,” they argued that the disparity in rules creates an unfair advantage for the likes of Google in the market. And that’s when I will at least give these critics some benefit of the doubt. Because as much as I may admire President Obama on a wide range of topics, his administration consistently tilted in Google’s favor across many areas of governance.

In his dissenting opinion regarding the FCC rules now on hold, Pai quoted the Electronic Privacy Information Center in rebuttal to Wheeler’s assertion that providers like Google and Facebook only see a “slice” of our data:

The FCC describes ISPs as the most significant component of online communications that poses the greatest threat to consumer privacy. This description is inconsistent with the reality of the online communications ecosystem. Internet users routinely shift from one ISP to another, as they move between home, office, mobile, and open WiFi services. However, all pathways lead to essentially one Internet search company and one social network company. Privacy rules for ISPs are important and necessary, but it is obvious that the more substantial threats for consumers are not the ISPs.

That certainly jibes with my own experience as a consumer. Regardless of how I get online—whether at home, through my mobile device, or via my local coffee house WiFi—all of my substantive activity occurs on web platforms.  So, yeah, I think privacy rules governing ISPs and “edge providers” should be both robust and consistent. But are they either robust or consistent?

That’s the nitty gritty we’re going have to try to follow in the coming months; and the task is not made easier when editorials fall into the trap of reporting this matter as Republican = corporate favoritism and Democrat = consumer protection.  When it comes to the interests of the Big Data companies, that narrative just does not hold up.  Chairman Pai’s core complaint about Wheeler’s FCC privacy rules is that the providers with the most detailed view of our data (.e.g. Google) would have far more lenient governance than the providers who have less insight into our private lives (e.g. Verizon). So, to the extent that Pai is correct in this assessment, I have to agree that consumer protection is best served by consistent rules governing the entire internet ecosystem.

Stay tuned. Much more to follow I’m sure.

VidAngel: A Litany of Copyright Defenses

Photo by rootstocks

VidAngel offers what is functionally a video-on-demand (VOD) service plus “filtering” for viewers who want to see mainstream fare with certain naughty bits—sex, foul language, violence, etc.—removed. To provide this service, though, VidAngel allegedly violates the copyright owners’ exclusive right of reproduction and public performance, as well as Section 1201 of the DMCA prohibiting circumvention of technical protection measures (TPM) used to encrypt DVDs.

Movie studios Disney, 20th Century Fox, Lucasfilms, and Warner Bros. sued VidAngel, and in December 2016, the District Court for the Central District of California issued an injunction, halting the defendant’s operations. The judge’s opinions in the order state that all arguments favor the plaintiffs’ likelihood of success on the merits.  VidAngel has appealed the injunction to the 9th Circuit where briefs were filed last week.

The Family Home Movie Act (2005)

As a frame of reference, if a consumer owns a feature film on DVD, a statute called the Family Home Movie Act (FMA), allows the use of technological means (e.g. a product called ClearPlay) to make limited portions of his disk “imperceptible” while playing it in a private viewing situation.  Designed for audiences who want to “filter” out scenes containing the aforementioned naughty bits, the FMA affords the viewer a limited right to use technology to achieve this “filtering” without infringing copyright.  (As a no-tech option, there is of course no limit to the amount a viewer may close his eyes, plug his ears, and sing La-La-La during those portions of a movie he finds offensive.)

We can debate whether or not “filtering” motion pictures is truly a right, but there is apparently enough of a market that wants to “filter” that the subject has at least been an issue.  For instance, the Directors Guild of America (DGA) was opposed to filtering technology like ClearPlay, claiming that editing the films violates the right of the author to disseminate a work as he sees fit, which is certainly true. So, the FMA was enacted as a legal compromise, written very narrowly to provide the home viewer with the ability to “make imperceptible” limited portions of legally-acquired DVDs.

Enter VidAngel

The CEO of VidAngel, Neal Harmon, reportedly grew up in a household where, for religious reasons, he was not allowed to watch a number of mainstream movies.  Remembering what it felt like to be culturally out of the loop, Harmon founded VidAngel in Utah in 2013—it is now headquartered in Silicon Valley—as a way to provide families in similar circumstances with a solution.  According to testimony, VidAngel approached the studios with their model; and the studios were not interested. There are any number of reasons why a film producer will be opposed to enterprise-scale “filtering” by a third party, and the studios were certainly under no obligation to engage in that endeavor.

Deciding to forge ahead, VidAngel developed a convoluted workflow and business model based on a bold legal assumption that the FMA allows the company to provide an unlicensed video-on-demand and “filtering” service while avoiding liability for infringement. The overly-complex model looks a lot like what it is:  an attorney’s Rube Goldberg attempt to circumvent copyright law. And although VidAngel wants to make this a story about the right of consumers to “filter,” launching the hashtag campaign #savefiltering, that debate has little bearing on the infringing nature of the VidAngel enterprise.

VidAngel’s Model/Workflow

VidAngel buys feature film DVDs, which they decrypt, digitize, organize, and store as “segments” that are tagged to facilitate “filtering” by pre-selected criteria (e.g. blasphemy). A customer uses the VidAngel app—available on Roku, AppleTV, Chromecast, etc.—to select among the criteria and create his “filtering” preferences.  Next, the customer “buys” a DVD from VidAngel’s inventory for $20—let’s say 1988’s Working Girl. But having filtered for nudity, he does not want to see Melanie Griffith vacuuming topless. (There’s a cleanliness next to godliness joke in there, but I’ll let it go.)

Most of the time, the customer doesn’t choose to receive the DVD he “bought”—it does still have the unwanted scenes, after all—but instead, he streams Working Girl, sans topless vacuuming scene, and then “sells” the DVD back to VidAngel for a refund minus $1/day for a standard-def stream ($2/day for high def). Note, the DVD never changed hands; VidAngel “held the DVD” for the customer during the “buy-sellback” interval.

According to VidAngel’s testimony, 80% of the disks are “sold back,” so if this sounds like an absurdly complicated way to operate what is primarily a VOD service, that’s because it is overly complicated—and quite on purpose. Because what VidAngel is counting on is that during the brief period when the customer “owns” the DVD, that customer may then legally “direct” VidAngel to perform the same function he would otherwise be allowed to perform at home under the FMA.  That’s their theory anyway.

The fact that VidAngel markets its service as providing streaming for “as low as $1” belies its claim as a “reseller” of disks; and the fact that 80% of its customers “sell back” the DVDs seems  more than sufficient evidence to reasonably describe the core business as video-on-demand. But even if a court might agree that a VidAngel customer is temporarily the “owner” of the DVD, this should have no bearing in assessing VidAngel’s infringing activities; and their own workflow makes this clear.

VidAngel Infringes Before a Customer Exists

In order to prepare files for “filtered” viewing, the company has to decrypt a DVD and then make and store copies of entire films on its servers (i.e. in fixed form). For practical reasons, the company must perform these two infringing activities prior to any customer “buying” any disks. There is simply no other way to organize the workflow and provide the service they offer. Because the infringement against the owners’ right of reproduction, and violation of the DMCA, occurs before a customer becomes part of the workflow, VidAngel’s claim that it is shielded by the FMA on the grounds that they “filter” at the direction of a DVD owner is simply impossible. The only owner of the DVD at the time when decryption and copying are performed is VidAngel. Moreover, VidAngel has predetermined a set of criteria for “filtering,” and no matter how many possible permutations of a given film this may produce, the company has still acted to create a finite set of “filters” it makes available to a prospective DVD owner rather than its claim to “filter” at the direction of an actual DVD owner.

Next, when VidAngel streams a movie, this constitutes a public performance in violation of another exclusive right protected by copyright.  Again, VidAngel puts its faith in the FMA, arguing that because they only stream a “filtered” version of a movie to the customer during the period when the customer “owns” the DVD, they are not publicly performing any more than if the customer himself were to engage in the same function at home using a legal “filtering” technology as permitted by the FMA.

Here, VidAngel appeals to the “spirit” of the FMA and not the statute, implying that Congress believed generally in the principle of “filtering” when it wrote the law. Whether this is true of Congress or not, VidAngel is asking the court for an extremely broad (dare I say, leap of faith?) interpretation of narrowly written legislation that does not allow a party doing the “filtering” to publicly perform a film beyond the confines of ordinary private viewing.   The FMA simply does not anticipate a model anything like VidAngel, which is exactly why the company is straining to create its own loophole in the law with its over-complicated pretense of “selling and buying back” DVDs.

VidAngel Claims Fair Use

Finally, VidAngel appeals to fair use doctrine, claiming that the “filtered” versions are “transformative” under the first prong, that the service is not a substitute under the third prong, and that the service does not create potential market harm under the fourth prong.  I suspect the appellate court will agree with the district court, which found that the fair use test favored the plaintiffs across the board; but it was the judge’s response to VidAngel’s fair use claim under the fourth prong that I find particularly revealing about the rationales being applied in VidAngel’s defense.  The court states:

“VidAngel attempts to support their arguments by offering customer survey results that indicate that over 51% of VidAngel customers would not watch their offerings without filtering. The survey results are ultimately detrimental to VidAngel’s arguments. The fact that 49% of VidAngel’s customers would view movies without filters shows that VidAngel’s service does serve as an effective substitute for Plaintiff’s unfiltered works, for approximately half of VidAngels users.”

Not that attorneys for VidAngel aren’t on the ball, but they actually presented evidence to indicate that nearly half of VidAngel’s customers may be poached from the potential customer base of the rightful owners of the works.  That certainly seems like a good way to fail on the fourth prong of the fair use test, but the self-defeating oddness of this argument is consistent with the major theme running through this entire case:  that the overly-complex design of the VidAngel model reveals a strenuous, ham-handed, effort to thread the enterprise through the legal boundaries of copyright.  It is little surprise that the lower court found all of VidAngel’s arguments untenable, and it is hard to imagine that the appeals court will not sustain the injunction.

A Moral Enterprise?

Advertised in their promo video as “Movie Heaven,” VidAngel presumably targets a market with sincerely held religious beliefs, even though the first laws they seem to have overlooked are Commandments #8 & 10: Thou shalt not steal, and Thou shalt not covet, respectively.  For all the assumed piety in VidAngel’s “filtering” crusade, the company clearly feels no sense of moral conflict about its for-profit, unlicensed exploitation of the thousands of people whose labor makes the movies. In fact, it is notable that among the content a VidAngel subscriber may have “filtered” from viewing is the end credits listing the names of all those workers.  I’ll leave the moral rationalization of that to the operators of VidAngel themselves, but as far as the copyright implications go, it seems like they don’t know what the H-E-double-hockey-sticks they’re doing.