Law in the Internet Society
It is strongly recommended that you include your outline in the body of your essay by using the outline as section titles. The headings below are there to remind you how section and subsection titles are formatted.

Facial Recognition and the Self-Sabotaging First Amendment

-- By JohnClayton - 20 Nov 2020

The First Amendment cannot sabotage itself. And yet that, in a sense, is what Clearview AI is asking a court to hold. Faced with a lawsuit for violating an Illinois privacy law, the maker of a controversial facial recognition app now claims the First Amendment safeguards it from all but the most narrowly tailored regulations. Clearview (so it argues) is no different than a news publisher: It gathers, synthesizes, and delivers information to its subscribers. Never mind that to fuel its platform, Clearview scraped without consent a purported three billion images from across the web—creating biometric faceprints of millions of unwitting individuals. But because the First Amendment protects the “the creation and dissemination of information,” Clearview’s activities cannot be proscribed.

This position is, I believe, untenable: the First Amendment cannot obliterate the privacy that the freedom of speech depends upon. But it is unsurprising that Clearview would make such a contention. The First Amendment has become a deregulatory tool, invoked to strike down laws regulating everything from traffic signs to the collection of health data. It will almost certainly loom over any comprehensive attempt to police the tech platforms; indeed, ISPs are already wielding the First Amendment against state privacy laws. Clearview’s case is thus both representative and uniquely disturbing. Few technologies are as overtly privacy-destroying as facial recognition. If free speech requires free thought, and free thought requires privacy, then it follows that Clearview’s argument cannot prevail.

Free Speech and the End of Anonymity

Over 130 years ago, Louis Brandeis and Sam Warren wrote of the right to privacy as “part of the more general right to the immunity of the person.” The “right to be left alone,” in their view, anteceded one’s ability to think and act autonomously. The freedom of speech, in turn, flows from this ability to think for ourselves. “If we did not regard the autonomy of the individual mind as important,” Seana Valentine Shiffrin writes, “it is hard to see why we would value its expression or outputs.” The instrumentalist value of speech in advancing self-government—long the centerpiece of First Amendment theory—correlates with the freedom of the minds from which the speech originates. Privacy, autonomy and the freedom of speech are inescapably linked.

The exigencies of the data economy jeopardize our privacy and autonomy. Facial recognition tools, in particular, cast doubt upon the possibility of anonymity. Hundreds of police departments across the United States have rushed to license Clearview’s app. And Clearview’s code reportedly allows it to be paired with augmented reality glasses that could identify individuals in real time. The ability to be alone in public may soon be a relic.

The obliteration of physical anonymity imperils free speech and thought. Facial recognition has increasingly been deployed to identify protesters, underscoring how these tools may deter core expressive and associational conduct. Online expression will also be chilled if one’s photos can be scraped by any developer for biometric indexing. Technology may offer some reprieve. New software can imperceptibly change images to make the individuals in them unrecognizable to machine-learning programs. But this does little to address the three billion photos already in Clearview’s database.

Despite all this—the appropriation of user photos without permission; the purposeful erasure of anonymity in pursuit of profit—it is Clearview, not those swept into its database, that seeks the First Amendment’s shelter.

Clearing a Path for Political Change

Legislators have increasingly acted to restrict facial-recognition technologies. Numerous cities have banned their use. A Senate bill would prohibit the use of biometric surveillance by the federal government and withhold grants from states that use facial recognition. Other laws, like the Illinois statute at issue in Clearview’s case, prohibit the private use of biometric data without consent—potentially cutting off an essential data source for developers.

The First Amendment should not sabotage such privacy laws, merely because regulated corporate actors trade in something that looks like “speech.” I do not argue that Clearview’s app—or other facial recognition tools, or tech platform algorithms in general—are beyond the scope of the First Amendment. The act of gathering and creating information, be it via a news article or a faceprint, would seem to fit under the broad umbrella of speech. So does Clearview’s program code; indeed, a contrary conclusion would expose privacy-enhancing tools like encryption code to intrusive government regulations.

First Amendment coverage, however, does not equate to immunity from regulation. Courts need not restrict the amendment’s scope or necessarily construct new doctrinal tests (beyond the strict or intermediate scrutiny formulas) to prevent the First Amendment from destroying the privacy its continued vitality requires. Rather, my critique is normative. In considering current and future challenges to privacy laws like Illinois’s, courts cannot blind themselves to the fundamental relationship between secrecy, anonymity, and free speech. They can follow the lead of the Supreme Court, which has championed anonymity in political speech a “shield from … tyranny” that “exemplifies the purpose behind the Bill of Rights.”

In a case like Clearview’s, then, the question backgrounding any analysis must be this: Does the private company’s “speech” imperil the foundational principles—autonomy and privacy—that the First Amendment sits atop? If the answer is yes, then courts must uphold any reasonable law meant to protect citizens against the encroachment of private power.

Conclusion: Beyond Facial Recognition

The dilemma of the self-sabotaging First Amendment is not limited to facial recognition. Any law that targets how tech platforms share our data will likely encounter similar challenges. Ditto for attempts to regulate how social media algorithms exploit our attention and chip away at our ability to think freely. To create an internet that nurtures human thought, we must tame an increasingly inegalitarian First Amendment. We can start by getting back to First Amendment first principles: that is, understanding that the freedom of speech that cultivates democracy and individual flourishing cannot exist without the freedom of thought. And the freedom of thought cannot exist without privacy.

On this logic, political speech design to "imperil the foundational principles ... that the First Amendment sits atop" should also no longer enjoy the protection of strict scrutiny. But that's not "the traditions of our people and our law," as it were.

You have seen clearly the tension between protecting the rights to learn, to think, and to speak and the protection of individual privacy against people who learn, think, and sell the results of their thinking about others. Whether the particular case of Clearview AI deserves the illustrative prominence you give it depends not on whether they are sleazy, which they are, but on how clearly the principles in conflict can be seen from a single illustration.

The route to improvement here, I think, is to honor fully the complexity of the phenomenon. As you rightly say, the US First Amendment is a deregulatory principle: that's what happens when Congress is told to make no law, while other societies are busily making laws that Congress and the State legislatures clearly cannot make.

But perhaps the issue isn't whether the First Amendment over-protects freedom of thought against privacy regulation. Perhaps the question is instead how to make privacy regulation consistent with the overarching commitment to freedom of thought. That would be the problem, at any rate, for those whose commitments to both are very strong, and who therefore are likely to reject an approach that requires them to be in conflict unnecessarily. That's my situation, just to take one guy at random, so in the next course I have to do that work. Wherever you want to take this essay, I hope that I will be able to respond to it, and to keep the conversation growing, in Part One of CPC.

You are entitled to restrict access to your paper if you want to. But we all derive immense benefit from reading one another's work, and I hope you won't feel the need unless the subject matter is personal and its disclosure would be harmful or undesirable. To restrict access to your paper simply delete the "#" character on the next two lines:

Note: TWiki has strict formatting rules for preference declarations. Make sure you preserve the three spaces, asterisk, and extra space at the beginning of these lines. If you wish to give access to any other users simply add them to the comma separated ALLOWTOPICVIEW list.


Webs Webs

r3 - 31 Dec 2020 - 13:50:40 - EbenMoglen
This site is powered by the TWiki collaboration platform.
All material on this collaboration platform is the property of the contributing authors.
All material marked as authored by Eben Moglen is available under the license terms CC-BY-SA version 4.
Syndicate this site RSSATOM