Saturday, March 16, 2013

Santa Clara Internet Law WIP part 4

Speech #3

Jane Bambauer / Is Data Speech?

Any time the gov’t prevents you from knowing something for the purpose of preventing you knowing it, the 1A is implicated: a negative right to create knowledge.

Definitions: Intentionally recorded information is data; information is an objective representation of something that happened. Examples: handwritten notes, photos, server data.

How can server data be speech if no one has examined the data? When gov’t regulates data, legal rules cluster around two lifecycle phases: data in transmission (diagram shows stick figure requesting data and other stick figure saying “sure,” which is clever and not entirely reflective of the less explicit transactions that gov’t tends to regulate) or data being collected.  This data is so similar to other forms of celebrated speech that it isn’t a close issue. Data in transmission is speech in the case law.

What about collection? Strong strand of precedent treats collection as nonspeech conduct.  Most scholars assume that collection can be regulated without 1A problems, and Bambauer was one of them until recently.  On reflection: Dietemann v. Time, Inc., involving a quack exposed by Life magazine; reporters surreptitiously recorded what was going on.  9th Cir. held reporters had 1A right to report and observe, and even to publish the photos, but no 1A protection for surreptitious recording of the photos/sound. That seems bizarre on reflection, though it’s how the cases lie.  How can the 1A shield the reporters’ minds and notes but not new tech that records and preserves in a more reliable way? And how can transmission be speech while the creation of the transmitted thing isn’t speech?

When the gov’t isn’t trying to prevent dissemination, it’s trying to prevent creation of knowledge, which is troubling.  Her position: A right to create knowledge: freedom from state action that purposefully interferes w/knowledge.  Does the gov’t intentionally/purposefully interfere w/the creation of knowledge? If yes, scrutiny follows.

Borrows from a thought-centric framework, Seana Shiffrin.  Hasn’t been applied to data privacy yet. Gov’t shouldn’t interfere with the creation of new thoughts; free speech makes no sense without that predicate right. Data has uprooted bad ideas, like the idea that ulcers were caused by stress. Moneyball taught us that scouts don’t have an “eye” for talent. The belief that homo sapiens killed the other hominids has been refuted by DNA evidence.  Public good theories: self-interested demands to limit speech tend to overwhelm the public good. See that in data privacy—earlier article on this.

Deliberative democracy: others point to cases in which speech relating to public concern gets the most scrutiny; protection drops off.  This isn’t a conception of the 1A she can get behind; limiting the scope of speech rights gives the gov’t the opportunity to decide what we should be concerned about. Even Brandeis wouldn’t have had such a narrow view of the 1A—was a great follower of John Dewey and thought that we all need customized, private self-education to be well-informed.  (Was that privacy tort Brandeis or some other Brandeis?)  Lending data seemed private but is now central to our understanding of what happened in the housing crisis.

Self-determination/autonomy: privacy of course often seeks to promote exactly these same interest by giving us safety/security to experiment.  Julie Cohen’s vision for autonomy through privacy is here.  She is very concerned with these autonomy interests.

What the gov’t can do under her thesis: can protect important interests in seclusion and confidentiality; prohibit certain uses of information even when a company knows something about its consumers—if there’s something wrong with a particular transaction, like discrimination.

Q: Dietemann: does the case come out differently if he was in an office rather than his house?  Students think so. Is that right?  If so, isn’t it less of a data collection case?

A: maybe, but the reason it’s right is that houses are related to seclusion interests. It’s not that the 1A doesn’t apply but that it might be outweighed.

Q: are you arguing to only protect the collection of the data or is it more important to disseminate it?

A: both important, independently

RT: If data is speech, then disclosure of data is forced speech.  This has significant implications.  For example, from the paper: “[The person whose data is collected], too, is free to gather data and make judgments about the best creditor or best prospective mate.”  The rich as well as the poor are free to sleep under bridges! Note also that this is not true: the corporation will be able to protect its privacy, and trade secrets, under the very right not to disclose that the claim of data as speech enables.

A: wants to think about compelled speech more; the doctrine is a mess. Can’t defend trade secret law; we need to fix that too.

RT: But fixing one and not the other has distributional consequences. Further from the paper: “generalizing based on categories … Without it, we must choose either arbitrariness or unconscious, unaccountable generalizations. There are no other options.” I worry about dichotomies presented this way. I thought many complaints about “generalizing based on categories” went to the way in which such generalizations are, in practice, arbitrary (redlining) and unaccountable (the camera points one way).  Accountability is not a predicate for First Amendment protection by Bambauer’s theory.  [Bambauer also posits that Big Data will have to be good because otherwise it won’t work well—I have some questions about how to apply that promise to, say, the models that brought us securitized subprime lending and the resulting AAA-rated CDOs.]

A: we have lots of disclosure laws that pass scrutiny. Compelled speech doesn’t need to have the same breadth that the negative right against interference w/knowledge creation would.

Q: if you support the intrusion claim, don’t you have to distinguish gathering and dissemination?

A: we have confidentiality rules and torts that do survive scrutiny; we’d do the same thing with intrusion.

Q: we’d have to redefine accounting law—a lot of laws about accounting; securities disclosure.  [Making you create data is interference with knowledge creation.] Is attorney-client privilege constitutional?  Do we do strict scrutiny on each to keep them in place?  Right now, a lawyer selling her client list would be a major violation of the law, as would a doctor selling her list.

A: if it’s speech, we do the scrutiny, and a compelling interest can win.

Q: but dozens of laws are on the rocks; the implications are far broader than privacy laws.  Underselling how sweeping this is.

A: links back to compelled speech.  But no one’s telling you you can’t do wacky accounting if you want.  (Though you will go to jail unless you use GAAP in various disclosures, at least if we put people in jail for accounting fraud any more.)  Compelled speech doesn’t mean that the bookkeeper can’t also do something creative with the numbers. 

Q: but that would often be fraud if it doesn’t reflect reality.

Q: maybe that would fall under prohibited uses of data.

Q: data isn’t mostly about privacy. Data is pervasive; business law is hugely about data. If narrow tailoring is required for any regulation of data, that’s superbroad.

Wu: are you requiring strict scrutiny?

A: no.

Wu: well, the other moving part is the level of scrutiny.  You may be ending up with much more scrutiny than you intend—think about affirmatively suggesting forms of scrutiny that don’t exist under 1A law.  [Of course, Fred Schauer does the foundational work on this issue of boundaries and expansion, and Vincent Blasi’s pathological perspective is relevant as well.]

[So does this theory of what’s subject to strict scrutiny mean that bans on marijuana based, at least in part, on claims that pot makes kids willing to try other drugs—that is, it creates new thoughts about other drugs—are subject to scrutiny? I agree with Wu that level of scrutiny needs attention; the easy answer is that the harms of marijuana use are sufficient even without that knowledge-based justification.  But it’s generally possible to regulate from a harm perspective.  E.g., from the paper: “Even hostile work environment claims based on adverse insults and aggressive or provocative speech attempt to regulate based on the effects on the victim rather than on the immorality of the speech itself.” I don’t find this distinction persuasive, once you change the strawman word “immorality” to “discriminatory content”—I had this discussion with Felix Wu before.]

Anupam Chander and Uyen P. Le / The Free Speech Foundations Of Silicon Valley

Le: Free speech helped create cyberlaw; the internet helped realize the vision of the First Amendment.  The 1A as industrial policy.  We aren’t making the case for a new Lochnerism. By focusing on expressive core of internet, we hope to restrain any libertarian vision.  Paper shows how the 1A helped configure American cyberlaw: CDA §230, DMCA, and a bit on privacy. We look at current threats.  SOPA/PIPA, operation In Our Sites where civil forfeiture/seizure were used against alleged copyright infringement, turning into prior restraints.  Global efforts to control the internet through the TPP, which also permits seizure/forfeiture and expands criminal copyright law/secondary liability.  Also the ITU, which recently considered proposals implicating speech—ability to identify speakers; relevance of national security; possibility of charging for internet traffic so that the sender pays. These aren’t domestic laws, but the US could be given int’l obligations.

Chander: the 1A isn’t a trump. But its role as normative force hasn’t been recognized throughout cyberlaw.  We’re also saying that it’s not the lonely blogger that we first envision, but actually intermediaries, whose role hasn’t been recognized. Readers are important, and the ability to connect people comes from intermediaries.  Disintermediation was thought to be the role of the internet, but it turns out it’s really intermediaries, and cyberlaw is configured in recognition of their role, just as NYT v. Sullivan recognized the role of speech intermediaries.  (See Marvin Ammori.)

Andrew Gilden: how does this account deal with Eldred and Golan?

Chander: Lessig et al. argue that speech interests should cause us to get rid of these extensions. SCt decided that the 1A wasn’t really at stake.  We have a piece on Golan—much of the activism was internet-driven; the relationship to those cases is one we should think more about. Not ultimately central to what was happening on the internet.  Activism bait, but not central to business models.

RT: Talk to me about what Evgeny Morozov would say about internet exceptionalism and people just wanting to keep government out of their business, using free speech as a bludgeon.

Chander: has a paper on Morozov’s last book—Jasmine Revolutions.  Morozov says people use the internet for porn; they get distracted and don’t use it for more ennobling pursuits. Extreme examples.  He says that if Burmese monks had the internet, they would’ve been less active protestors.  His skeptical view is overdramatized. We’re arguing not for exceptionalism but for realizing the 1A’s promise, one never realized by handbills or by the NYT.  Internet doesn’t deserve unique protections even though we should recognize its unique role.

John Tehranian & David Levine / It’ll Break Your Heart Every Time: Baseball, the Internet and Romanticism

Tehranian: When you look at Flood v. Kuhn, SCt gives absurdly detailed history of the game; case involved a player who didn’t want to be traded. SCt had to decide whether baseball continued to have an antitrust exemption. Yet the decision doesn’t discuss the history of labor relations in the game, or race relations, setting up the denial of his claim.  Delegitimizes Flood through its analysis, while ignoring the change in “commerce” that had occurred during the period; the old precedent about the antitrust exemption came from a very different constitutional order, but the Court didn’t recognize that. Highlighted his high salary, not his claims for the exploitation of his labor.

Levine: same romanticism in Citizens United. Court has equally romantic statements about the internet as in baseball. Core of majority opinion waxes poetic about internet’s ability to balance speech playing field right now.  (So Evgeny Morozov does have a point, eh?)  That’s not true given the realities of political campaigning. The Court assumes a level playing field—a blog post can be as effective as a corporate media buy. They say they don’t want to hurt the progress of the internet by denying it the ability to seek its fruition by balancing corporate speech. Prompt disclosure provides citizens with the information needed to hold corporations and politicians accountable. As someone who writes about gov’t and private secrecy, he’s highly dubious.

Both baseball and internet are held up as great for the country, but SCt views them through rose-colored glasses.

Tehranian: romanticism invites a failure of imagination. The Court held that baseball absolutely needed the reserve clause, preventing free agency, and therefore needed the antitrust exemption to protect the integrity of the game. Proven unequivocally untrue when baseball was forced into free agency. During the 1980s, baseball had more competitive balance than in decades; revenues exploded. The romanticization/desire not to change can lead to blind adherence to precedent or flagrant disregard of precedent.

Q: is this something you expect to find whenever the SCt does something flagrantly wrong?  You think the SCt is wrong and reasoning screwily; aren’t those likely to be correlated?  (Compare Dan Kahan’s Harvard Forward and the responses to it.)

Levine: courts tend to play around with claims about the internet; maybe these results would happen anyway. But to the extent we have data available and used in ways that are irrelevant or ignored, we have a parallel.

Tehranian: true, could argue this whenever you think the Court is wrong. Interestingly, Flook’s deference contrasts with Roe v. Wade and other contemporaneous decisions—deference is strange.

Gilden: romanticism may be inherent to how you make arguments as lawyer and judge—offer a steady state that needs protection. It’s indeterminate: you can make romantic arguments on both sides of same-sex marriage.

Tehranian: fair enough—need to ferret out strategic use to sell a decision v. an actual explanation. Flook involved Justice Blackmun really, really believing in the narrative he offered.

Luna: to what extent are you observing the cause or an effect covering for another cause that is less easily observable? Would be useful if you specified a bit more what kinds of values you would include as romanticism.  Are you thinking of particular ideologies?  Conservatism?  Tradition?

Ochoa: the story of the Founders is romanticism—any reference to Founders’ intent could be read that way.  Skeptical that you’ve identified something unique; they ignore data all the time.

Joseph A. Tomain / Online Privacy: First Amendment Analysis of an Opt-In Regime

Problem with proliferation of ads and marketing on and offline.  Current proposals—European Data Protection regulation; FTC proposals.

Are current online data collection practices misleading?  He thinks so. Is the gov’t interest substantial?  Yes, individual privacy.  An opt-in requirement directly advances the gov’t interest, and not more extensive than necessary since online data collection is still permissible.

Sorrell: law burdens disfavored speech by disfavored speakers—that can’t possibly be right if commercial speech doctrine survives. Court said that detailers/pharmacos have a right to communicate in an effective and informative manner—but does that give you a right to reach people you don’t otherwise reach?  Not workable.  Also, are these communications even informative? Tamara Piety’s Brandishing the First Amendment distinguishes between information and propaganda.  Default rule should comport with the preferences of most people—opt-in before data collected. Prior cases have given surprisingly little attention to opt-out v. opt-in—opt-out was more historical accident than purposive.

Sorrell could be impediment to opt-in regime, but shouldn’t be under Knox v. SEIU which allows it.

Ochoa: why is this speech?  Facts are in the public domain for copyright, anyone can copy them.  Is collecting facts speech?

A: he thinks so, when Google uses my search to present ads.

Ochoa: if collecting data isn’t speech, you don’t need a 1A analysis.

Bambauer: this is automatic, not really “collection.” Opt-out is really data destruction.  Not about collection, but data that is use-restricted.

Q: information is not itself harmful—people can perceive their best informations if they’re well informed.  Sorrell says it’s about the benefit of consumers.

A: Va. Pharmacy was about drug price ad. Sorrell wasn’t about consumers; it was about providing info to marketers.

Q: but Sorrell was premised on the idea that information (not the data transmitted between the parties, but rather other information related to their data) would get to consumers. The real issue is the ability of consumers to make choices; you can constrain the availability of information at the point of collection.

A: difference between info and propaganda: marketers aren’t telling me stuff in my best interests.

Q: so you’re comfortable w/gov’t drawing line between persuasion and propaganda?

A: I have problems with large private actors making those decisions too.

RT: Re: misleadingness—if misleading, can be banned outright; careful about the standard for what counts as misleading. Also note Lauren Willis’s important work on companies securing opt-ins through clever design.

Q: Europeans don’t allow monopolists or people w/market power to count as opt-in consent.  But DC and 10th Circuits go in opposite ways on opt-ins; DC in Transunion v. FTC held that opt-in was just fine, while 10th held it was unconstitutional.

Q: Consider costs to people who don’t value their privacy.

No comments: