First Amendment Opportunism: Heads, Facebook Wins; Tails, We Lose

Frank Pasquale

Oft-discussed at Yale's black box conference in April, Facebook bias is now a national news item: a Senate investigation is on the way. Michael Nunez has interviewed former curators at the media/platform/AI/VR behemoth, who say conservative news (and news critical of Facebook itself) was routinely suppressed. Experts are developing proposals to ensure fairer and more transparent social media.

I think these proposals are, by and large, commendable, but I worry that they will be deflected by glib arguments. The elite press has all too often swallowed Google’s characterizations of the “right to be forgotten” as an affront to basic principles of freedom of expression. They will probably be receptive to Facebook's claims that it is little more than a digital town crier, which must be protected from the heavy hand of the state.

So, in the spirit of Chris Hoofnagle’s Denialist Deck of Cards, let’s review some of the analogies most likely to be deployed by Silicon Valley lobbyists, and why they are unconvincing, contradictory, or both.

"The algorithm is too complicated for anyone to understand--how can you regulate it?": This is a classic move for large internet platforms. “How could we be biased? Even we don’t understand what’s going on:” I’ve heard this from engineers, journalists, and computer scientists. But even if the data processing is too complex to understand, we can still regulate what data goes into platforms’ algorithms, and how the outputs are used. Moreover, it's not even relevant here.

"We have a clear point of view—we’re speakers with full First Amendment protection": The “complicated algo” line is not likely to work in this case, because Nunez’s report features curators who explicitly state they were told to inject bias into trending topics. So expect a diametrically opposed line: rather than characterizing trending topics as too complex for anyone to understand, now friends of Facebook will say the firm expresses a clear political point of view. “We’re a liberal company—you can’t force us to prioritize conservative news. We’re the New York Times of the 21st century!”

"But we’re not responsible for defamation or IP infringement--we're a conduit, not content": Forget, for now, whether an aggregator with enough financial clout to buy most mainstream newspapers (or selectively pick off uncooperative “partners” by reducing traffic to them) can credibly analogize itself to the Times, or the Miami Herald. Just consider Facebook’s responses to “freebooting” (massive-scale infringement of copyrighted work). If the New York Times republished, verbatim, an article from the New York Post, it could be sued for copyright infringement. But Facebook is immune from suit, under the DMCA, as long as it puts some effort toward responding to take-down requests. The same asymmetry applies to defamation claims: a publisher can be put out of business thanks to a defamatory story, but intermediaries in the US hide behind the CDA, even if they enable the story to be the top result on a person's name for the rest of the person's life.

I am not commenting on the wisdom of the DMCA or CDA here; the problems of overenforced IP law, or collateral censorship, are real. I just want to convey the extent to which Silicon Valley giants’ success depends on clever adjustments of old legal principles that tend to leave them with the rights of legacy firms, but few of the responsibilities. While Google has, at least in part, recognized some responsibility by ending ads from payday lenders, Facebook is lagging. And it is all too prone to deploy expansive interpretations of the First Amendment to sandbag even the mildest proposals for transparency and accountability.

The moral force of "freedom of expression” is weakened when it's reduced to one more cliche in the Silicon Valley PR machine. Companies like Facebook are quick to take the credit when their platforms are part of movements, but trivialize user rights in their own governance. It is almost as if the platforms see themselves as virtual worlds, whose users have essentially accepted (via terms of service) near-absolute sovereignty of corporate rulers.

At their best, platforms recognize that such sovereignty comes with responsibilities as well as rights. The power to rank is the power to make certain public impressions permanent, and others fleeting. As platforms gain commercial, political, and cultural influence, they are the new kingmakers. The question now is whether the state (and companies themselves) will make these processes more comprehensible, fair, transparent, and open to critical analysis by the publics they affect.

When massive platforms combine the functions of conduits, content providers, and data brokers, analogies from old free expression cases quickly fall apart. Too many discussions of social networks and speech are nevertheless moored in murky doctrinal categories, reifications, and inapt historical analogies that do more to obscure than reveal the true stakes of disputes. It is time to think beyond the old categories and to develop a new way of balancing dominant platforms’ rights and responsibilities. Sometimes, that will require the translation of old principles of media regulation (like rules against stealth marketing and unfair/deceptive acts and practices) to new contexts. In other cases, litigation will be needed to stop dominant platforms from abusing their power online. Platforms should also acknowledge their de facto role as public forums, quasi-judicial law interpreters, and fiduciaries, even if they resist taking on all the de jure responsibilities such roles imply for older entities.
Posted 1:24 PM by Frank Pasquale [link]