Tuesday, February 18, 2014

HLR Free Speech Symposium: Jack Balkin, Old School/New School Speech Regulation

Speech presupposes an infrastructure. Sometimes obvious, sometimes not.  NYT isn’t simply the print on the page, it’s a staff and a building and printers and printers’ unions and delivery trucks and ad agencies and all the infrastructure that allows that to happen.  Public libraries; postage subsidies; many different things that make idea circulation possible. Many old chestnuts of 1A law are attacks on infrastructure: CIO v. Hague, attempt to deny CIO any place to meet in public.  Ban on handbills = attack on mode of expression. 

Distinction now between way in which speech was primarily regulated before the digital age and new ways of regulating.  Old school is still around; new school techniques are all aimed at digital infrastructure.  Hardware, institutions: payment systems, search engines.  Large set of practices and institutions. Old school target is speaker/publisher; new school target is elsewhere.  Old & new school supplement each other.  Edward Snowden: they bring down an airplane to try to catch his body; British version of NSA detains Glen Greenwald’s spouse as old-school courier; Chelsea Manning’s military imprisonment used to make an example. 

New school techniques: variations on the idea of prior restraints.  Collateral censorship.  When you have a few media companies, state can own them all; can license them; can use family/political ties to control them. Not that difficult for state to have continuous relationships with them—a global phenomenon. But what if you multiply the numbers? Traditional strategies of control have to be changed—so you go after platform/conduit instead.  Platform has different incentives and will overblock/censor the individual speaker.

Digital prior restraint. Not the Pentagon Papers, but prior restraint also covers licensing/control systems, many administrative.  Prior restraint is when the state restricts first, gives permission later. Deliberate overbreadth. Low visibility.  Shifts burden of inertia and action.  Force people to identify themselves, promoting self-surveillance.  Filtering: direct by gov’t or inducing private parties—deliberately overbroad, no hearing before the vact; very low visibility, shifts burden of inertia to you. Domain name seizure is the same way.

SOPA/PIPA were designed to reorient how IP rights were enforced online.  Ingenious new techniques for getting at pirates—many had digital prior restraint, such as AG’s ability to allege that a foreign infringing site (defined to include some nonforeign sites) exists and get an injunction.  AG could get ex parte order and go to search engines, advertisers, payment systems requiring them to block/not do business with the site.  Break the internet (don’t resolve the DNS): that’s a prior restraints. Private prior restraint: if someone alleged that a site was dedicated to the theft of IP, that private person could send a letter to advertisers and payment systems.  Ex parte and not even in front of a judge. 

McCarthyism: public/private cooperation to create blacklists. You can do this in the new school too. Your ISP/infrastructure provider can provide the gov’t with data access; gov’t can give them immunity for that.  Can give immunity for collateral censorship.  Can give immunity for blacklisting people w/whom you refuse to deal. 

New form of soft power, in which gov’t can’t actually go after Wikileaks for a number of reasons. But it can suggest to others that they not deal with Wikileaks.

Finally: national security letters.  Gov’t has to cooperate with/commandeer private structure of free speech to surveil—so infrastructure of free speech is merging with infrastructure of surveillance.  National security letter comes with a gag order. Not even before a judge ex parte: the gov’t just approves it through the FBI.

In the courts, it’s proven difficult to water this down. If you applied even Friedman v. Maryland (the rules for dirty pictures), it wouldn’t work because courts couldn’t handle the 1000s of NSLs each year. Most times Google and Yahoo don’t want you to know how many NSLs they receive, because that makes them look bad.  Also, NSLs largely affect overseas customers, which is bad business for them. 

Chilling effects: when you combine prior restraint with surveillance state, in most cases the gov’t would like its surveillance to be in the background; gov’t doesn’t want you to get anxious or stop talking or travelling with your GPS enabled phone.

Dawn Nunziato: Cooperation and informal controls are weakening traditional constraints.  Optimistic story to be told: Regarding prior restraints, 1A doctrine is still efficacious, esp. compared to Western democracies.  Our 1A precedent/values are increasingly informing int’l free speech protections and industry self-regulation—industry players are resisting some of the cooptation. Not focused on national security issues; more on content regulation.

CDT v. Pappert 2004: Penn. law requiring ISPs to block/take down content on/accessible through their servers, on receipt of ex parte judicial order on submissoin by AG of probable cause that content was child pornography, or on receipt of informal notice from AG that content was child porn without court order: if not taken down after court order, ISP could be held criminally liable. 

Danger of overblocking.  Despite informal nature of requests, despite illegality of child porn, despite role of private parties, despite probable cause and judicial determination, court found illegal prior restraint.  Ex parte/no requirement that publisher/distributor receive notice or have opportunity to be heard.

Informality of notice didn’t save the Act.  ISPs didn’t want a court order; would prefer informal notice, but those were still subject to constitutional scrutiny.  Earlier case: Bantam Books involved informal notices sent by admin agency to publishers with request for cooperation w/commission to protect use.  Still unconstitutional prior restraints if unaccompanied by requisite const’l safeguards.  Similar result where public library filtered to block access to child porn, porn, materials harmful to minors: Mainstream Loudon.  Even ALA v. US upholding mandatory filtering as condition for library funding rested on ability of adults to have filters removed, per Kennedy and Breyer (concurrence whose votes were necessary to result).

1A is still effective by comparison to other democracies. In UK, Internet Watch Foundation is private entity w/responsibility for maintaining blacklist of websites w/child porn, hard core/extreme porn; used to be responsible for racial hatred sites.  Since 1996; blacklists affect 99% of internet users, because ISPs are required to check against the blacklist. No judicial supervision, notice, or opportunity to be heard by affected websites or users.  David Cameron plans to impose mandatory nationwide family friendly filtering by default on all computers by end of 2014; Australia has also made efforts in this regard.  UK’s mandatory filtering blocked an award winning British sex education site, BishUK.com, as pornographic without notice.

US prior restraint law is influential—European Court of Human Rights is actively borrowing.  Yildirim v. Turkey, 2012: found violation of European Convention Art. 10 in Turkey’s blocking of Google Sites to block an anti-Turkish Google Sites page: borrowed Bantam Book’s language of prior restraints being presumptively invalid.  2011 report of UN Special Rapporteur on freedom of opinion and expression builds on key 1A elements.

Sullivan removes barriers that disproportionately discourage intermediaries from carrying others’ speech, but no ISP would risk Sullivan style liability--§230 was necessary.  Protects dontdatehimgirl.com, which allows critical statements about men; JuicyCampus.  Veseley v. Armslist, 2013: should protect arms sales, though several Congresspeople wrote that this wasn’t what they meant.  We don’t need to worry about this aspect of Sullivan’s legacy being underprotected or underenforced.

Private speech intermediaries are making efforts to resist cooptation and increase visibility and transparency.  E.g., Google’s resistance versus “Innocence of Muslims.” Industry group:  Commit to respect free expression.  Google’s Transparency Reports.  Microsoft and Yahoo have also begun to detail how they respond to gov’t requests to take down/censor content.  Twitter too publishes on Chilling Effects.

Yochai Benkler: Sullivan is a realist opinion. A venerable common law rule of private law is a rule that regulates speech.  Less virulent member of the line from the embarrassing Shelley v. Kraemer, which exploded state action.  (Carol Rose has a great article explaining why she thinks this isn’t accurate.) The thing we care about w/regard to speech isn’t just about gov’t officials deciding who should speak, but also about private law adjudication.  It’s that realism that suggests that the 1A may not be where the future of freedom of expression lies.

For example, there’s no prior restraint, but the single most linked-to post about the problems of SOPA/PIPA on the single most linked-to website (Techdirt) was unavailable for search on Google for a month because it was included on a list of 1000 links sent to Google as infringing under the DMCA.  Took weeks for Mike Masnick to figure this out.  That’s not about law and the 1A, just new school regulation.

All individual freedom can only take place in systems of affordances and constraint.  The phenomenon is ubiquitous.  If you read US v. Jones, you see that protection from state surveillance in the past is primarily cost/inefficiently in an environment not ubiquitously impregnated with sensors.  Both the new capabilities and the new constraints are functions of the disruption of the older model.

Crawford starts out by asking what the basis of free speech is. She focuses on the negative—use to constrain regulation—but underlying move is to retain architecture where bottleneck doesn’t have control. Common carriage becomes more important to free speech than the 1A, except insofar as 1A gets in the way.  Ammori: the source of freedom he describes is the organizational capacity of individuals acculturated in a certain mindset about the relationship between tech and freedom. This isn’t new: NYT and Wash. Post had the same model, allowing familes committed to model of professional journalism that couldn’t sustain the tyranny of the margin—used that model of journalism/professional norms/corporate law that allowed dual stock structure, not the 1A. 

US is influencing all the systems of public/private cooptation and collateral censorship through the USTR—trying to push them through in ACTA; requiring them in bilateral trade agreements.  SOPA/PIPA protest ripples through Europe/blocks ACTA.  Ammori played a central role in that network.  It’s that set of systems and their interaction that matters.

Rise of mobile communication as a risk to speech.  Speech isn’t protected so much as implemented these days. Open standards, open software, etc. Once you move to proprietary mobile handset, Apple can exclude a game that mocks its labor practices, an app for reading Wikileaks cables, etc.  Move from distributed storage to cloud = points of intervention and control for soft power.  This isn’t a self conscious gov’t move to change controls, though that exists too, but rather a change in the market dynamics of the tech that is shifting us for reasons of convenience and cost to an infrastructure that is vastly more regulable. That’s a central threat to freedom of expression without touching the 1A, and it’s more important as a threat than all the doctrinal questions, which are marginal.

A young woman who wants to express herself about sex through how she dresses: the threat of being killed is vastly more of a constraint on free speech than even an authoritarian gov’t, simply because of the closeness of surveillance. To say that there’s an improvement in free speech when the gov’t seems more isomorphic with the US seems to miss the point.

Theory of freedom generally, not just free speech. Overlapping systems of constraint and affordance that are necessary to allow us to act autonomously while also providing others with channels of power. Essence of freedom is capacity to read these systems and figure out how we can bob and weave between them.  That’s where freedom is. Any focus on one to exclusion of others risks blinding us.

Mark Tushnet: If this is about realism (Shelley v. Kraemer), many of these things seem doctrinally manageable. Getting back to Shelley on state action isn’t easy, but might be done.

Post-New Deal vision was a vision of statism, which is fine for Crawford and problematic w/r/t the surveillance state, but the realist resolution of that tension was to say, yes we are statists; the location for the discussion of these conflicts is through ordinary politics, not constitutional ajudication, and that includes politics about the structure of politics itself.

Q: wouldn’t we need to move away both from state action and the model of an individual challenging a specific gov’t action.

Balkin: yes. Many of these things could be dealt with by adjusting the prior restraint doctrine. But many of today’s problems are either problems of regulatory structure (if designed a certain way, no 1A problems arise) or a deeper set about the way in which the tech develops (no 1A right to not move to cloud based computing; no 1A right not to develop smart phones).  Congress isn’t even the right venue for that latter discussion.

Benkler: we tried to force courts to consider copyright/DMCA in 1A terms. The refusal had massive implications for the architecture of free speech, but the courts didn’t realize that.  If we accepted that the province of law is much narrower than initially thought, and that interaction between 1A and people is always playing a chess game on multiple boards, that’s a different kind of lawyering but doesn’t make the 1A irrelevant.

Benjamin: if we’d never had a 1A, would we think about Balkin’s examples differently?  Would we see them as ordinary political problems? How do the British think about these?

Balkin: distinguish between a text of the 1A and the power of judges to create implementing doctrine. Modern 1A doctrine is a product of the second half of the 20th century, created in short bursts in the 1940s and around 1976. Before that, the 1A is mostly norm-based, common law understandings. Fights over the post office, labor unions, abortion, etc. were political fights. We could talk about free speech traditions, but done through politics.  British too had a principle against prior restraint. We are returning to an earlier q—how to build out an infrastructure (like railroads and telegraphs) that supports free speech, without the courts.  Irony: Sullivan represents the moment at which we hand the 1A to the courts but once again it looks like we will have to resolve it through politics.

Mark Tushnet: we know what the British would do.  The common law set of entitlements created by background rules of contract, property, and tort can be adjusted in the ordinary way by thinking about the purposes sought to be achieved, subject to legislative revision.  Common carrier idea is exactly that.  Direct consideration of the background rule.

Q: if action is moving in direction of private restraints on speech being more concrete than public restraints, then we might move to legislation to force open private channels. How well is 1A doctrine suited to evaluating those sorts of situations, where gov’t is operating on a private party w/overarching purpose of enhancing speech?  (I wrote an article about this.)

Nunziato: that’s the discussion we had with Crawford. 

Benkler: greater capacity of some private and some public actors to use this system in ways that may disturb us. 

Balkin: so much more likely that 1A will be used to prevent reform in the next 20 years.

Q: ECHR is active here—privacy in the cloud.  Member States must adopt legislation to protect data in the cloud.  Net neutrality exists in the Netherlands, and the main argument there was Art. 10 of the ECHR, with positive obligation to protect freedoms.  Do you ever envision a future in which the traditional negative freedoms, which are very much limiting, will transition into a mix of negative and positive obligations?

Balkin: guesses that it could only happen through the development of judicial restraint.  There are some exceptions—public forum doctrine as a doctrine of access; limited rights of access to traditional court proceedings. But US free speech tradition has protected positive access through custom, legislation, and institutional reform.

Ammori: would our answer be different if we thought we could win more cases in the courts? He’s all in favor of forum shopping.

Balkin: courts that were more sympathetic would dismiss various challenges. 

Ammori: what about surveillance?

Balkin: is it possible to use negative traditions to protect the environment of free expression? Yes, we missed an opportunity and went instead only to the 4th Amendment, but we could use freedom of association to protect against surveillance.

Benkler: Lost disastrously in copyright in the White House/Congress in 1998; lost disastrously in the courts; then moved to culture/technology. The answer is consistent with a multisystem view of a substantive commitment to freedom and a pragmatic catholicism about which system to work in given openness at particular historical moment.

No comments: