Enrique Armijo / Kill Switches, Forum Doctrine, and the First Amendment’s Digital Future
What’s different about digital speech? Political scientist Larry Diamond: information and communication technology is liberation technology allowing citizens self-determination—samizdat/small self-publishing for change. Flip side: Sergey Brin, Google, more realistic: it turns out that you can put the genie back in the bottle—because speech is intermediated. In physical space, the First Amendment relies on the presumption that speakers and listeners share the same physical space; speech is limited by space and temporality, as in the public forum doctrine. The law looks at the space the speaker and listeners share and asks whether it’s public, who owns it, what its traditional uses are, whether it’s traditionally a public forum—if not the government can generally shut it down. Public forum as easement: public earns a right by speaking.
Citizen-to-citizen communication through tech—listeners don’t have to be in the same space or time as the speaker or as other listeners. Also illustrates intermediation: every point in the act of communication is facilitated by a third party. It can be a network, software provider, etc. We’ve overcome temporal and spatial limitations, only with intermediaries, and intermediaries have increased capacity for interference. Moreover, we lose public forums and any public easements in them.
We’ve already had ex post punishments, but tech expands the ways in which a government or private actor can interfere ex ante and made ex post more effective. Ex ante: kill switches, blocking individual sites/Twitter/etc.; website licensing requirements; restraings on private choices; state-built internet alternatives. Ex post: surveillance that deters speech; criminal punishment; countermeasures. Not a one way-ratchet towards freedom. Accounts not available in certain countries, e.g., Innocence of Muslims blocked in some countries, and Twitter accounts blocked in Germany. Google doesn’t owe you anything, including freedom: Siva Vaidhyanathan.
Collusion between private and government entities; shutdown of internet to Egypt and Libya.
In the US: BART cellphone service shutdown in order to prevent a public protest. This isn’t an equivalency argument, but it is an architecture argument: controlling the network is controlling speech. BART argued that it wasn’t a public forum—shutting off its repeaters didn’t implicate speech regardless of content, because a train platform isn’t a traditional public forum. But BART wasn’t just providing a physical space, it was providing a network. Even if the speaker is on the platform, the listeners may be elsewhere—their rights to receive information are compromised; their right to association too. Interference with speech affects both parties’ speech interests, but the public forum doctrine doesn’t recognize that.
Other doctrine is also unsupportive: US v. American Library Association: providing internet access doesn’t create a limited public forum, since it’s just for patrons’ research and therefore websites have no speech interests. Intent only goes one way. USPS v. Greenburgh, the mail: placing unmailable matter can be banned too.
What happens when these cases reach muni wi-fi? Public-private partnerships?
Think less like First Amendment scholars and more like engineers to protect speech. Common carriage principles: should serve all customers, carry all traffic, and discriminate against neither, regardless of identity of the customer or the content of the traffic. BART argued that it should be able to shut off the network if it knew that there was a cell-detonated bomb; but his rule would subject that decision to strict scrutiny. However, Congress has declined to create common carrier obligations.
Ai Weiwei said: “if the internet is uncontrollable, freedom will win.” But what happens if the internet is controllable?
Mark Blitz: there is a kind of intermediated speech that doesn’t fit the standard model—good old fashioned writing. Less intermediated than it used to be—didn’t need public forum doctrine to protect it. (Did need protection from antilittering ordinances etc.)
Larger question: who would have common carriage requirements? Does the Constitution command an answer? Most straightforward application is to ISPs stepping into role of phone company, but what about Twitter and FB, etc.? They have a stronger claim to be publishers as well as traffic carriers. How do you figure out who’s subject to the requirement?
A: could have content discrimination on the ends with an open channel, and he may be prepared to accept that.
Q: public forum doctrine is dead; may bolster argument with distributional argument—for people who can’t afford private access, gov’t provided access may be their only hope, so they’re disproportionately affected. Consider also Marsh v. Alabama, the company town case. Also consider the basis of the claim by the internet users against the state—listener-centric strands of First Amendment doctrine are poorly spelled out but might help. Also what happens when gov’t passes of enforcement to private parties, as with six strikes in copyright.
A: Agreed with the “dead” status, but listeners matter.
Q: in the 19th century, regulators had a lot of trouble figuring out what “discrimination” means for common carriers. What would it mean today?
Felix Wu: how would this apply to a shutdown of muni internet services? Could be characterized as nondiscriminatory regardless of the reason—or are they required to give a reason?
A: mine has two components, carry all messages and carry all messengers. If you shut down access for a content-based reason, that would be a problem.
Wu: but is that true for common carriage doctrine? If you disallow all packets, then what?
A: ties into a postal service case: the gov’t is free to have a service or not have it, but once it has a service it has to deliver any mail. (But it probably can’t shut the mail off during a political campaign and then turn the service back on.)
Q: but there wasn’t universal service until the late 19th c., just big city deliveries and then private contractors who’d deliver it to individuals or you’d go to the federal depot. Rural delivery was a lot longer.
Blitz: could borrow useful pieces from public forum doctrine—neutral time place and manner regulations ok, but non-neutral/selective enforcement would violate the Constitution. Can go beyond a very simple version of common carriage requirements.
Ira Steven Nathenson / Super-Intermediaries, Code, Human Rights
Spurred in part by Innocence of Muslims: with great power comes great responsibility; Google sometimes casts itself as the hero, but is also often cast as the villain. Recognize the importance of social media in spreading info through the Arab world—Twitter used by 1.3 million Arabs and FB by 43 million.
Super-intermediaries: a term used in banking law; his claims here are both descriptive and normative. Certain intermediaries have great power and thus special responsibilities to the public—does int’l human rights law help guide us on what those would be? Then tries to provide a taxonomic framework for which intermediaries count. Avoid express normative claims about what speech ought or ought not to be protected or how to balance claims under int’l human rights law. Does make claims about the process super-intermediaries should use and the transparency they should maintain about takedowns.
If we want to look at powerful intermediaries, look at social impact. User relationship (how interactive is it; how individualized is it; networkability)—treat Google search separately from YouTube; legal scrutiny; reputation (may have political power, even celebrity status, and hero/villain ambiguity).
Legal scrutiny: intermediaries that face a lot of private legal scrutiny, lawsuits from individuals, include Google, Amazon, FB—not so much for Craigslist or Wikipedia. Suits for antitrust, patent, copyright, TM, and defamation—all claims (except patent) likely to be asserted against an intermediary because of its intermediary status—the list is pretty much the same. Private extralegal scrutiny—hundreds of thousands of takedown notices also sent to really powerful intermediaries. His own ranking: FB, Twitter, YouTube, Google (search engine), then eBay, Amazon, Wikipedia, Craigslist, ISPs; Tumblr may be up and coming but isn’t there yet.
Superheroes: tropes of superheroes are superpowers, vulnerability, archenemies, even secret headquarters. Can find guidance there.
Innocence of Muslims. Cindy Lou Garcia’s acting was overdubbed with a question asking whether Mohammed was a child molester; many protests resulted. Obama administration asked for the video to be taken down, and Google refused in the US but blocked it in a number of countries. When searching YT Egypt site, he could easily search for it. What Google seems to be doing is geolocation, even though it’s using a country-specific TLD.
Universal Declaration of Human Rights etc. protect freedom of speech/expression, but there are significant tension: ICCPR says speech can be restricted to protect the rights or reputations of others, public health, public morals, etc. ICCPR 20(2), which the US didn’t join, says any advocacy of national, racial or religious hatred that constitutes incitement to discrimination, hostility or violence shall be prohibited by law.
Various ways that superintermediaries regulate—processes, algorithms, domain name system, geolocation. Ways of dealing with these situations: one way is the status quo. Tim Wu on community-based system a la Wikipedia, though that has significant downsides. The Innocence of Muslims page on Wikipedia in English is very different in French; this is not one global community but embodies different values, cultures, and speech choices. Pseudonymity is also allowed—risking trolling and shilling; whereas a real name requirement puts people living in repressive regimes at risk. Tribunals like a human rights commission? Code? Geolocation is difficult to opt out of. Content ID has a res, for IP, but can’t really do that to identify speech.
His own proposal: process principles. Transparency/public participation in creation of processes that can lead to removal of content. Transparency about the source of requests to remove expression. Transparency/public participation in decision-making for high-profile cases. Transparency for the outcome of requests to remove expression, as Google does. Processes for users in blocked regions to opt out?
Mark Lemley: There are efforts to voluntarily do a bunch of these things, like contributing to Chilling Effects. Google and now Twitter have started disclosure. Is this a set of best practices or a set of legal principles? If built into law, whose law?
A: might be product-specific—Picasa might not need these. For opt out, that’s a hard question. Geolocation lets Google decide to comply with local laws—should Google enable people to flout the laws of their countries? Difficult Q because he is punting on the speech issues. Maybe it’s just too big an issue.
Lemley: imagine me, secure in the US, saying “I want my video to show up in Egypt” v. an Egyptian saying “I want my video out there even though I know it’s banned here.”
A: his thought is that people should have a mechanism to get political/religious speech, whether it’s affirmative provision or not blocking attempts to get around geolocation.
RT: as a comic book fan, I want to make a suggestion about the paper’s discussion of secret identities: a better analogy would be the use of tax havens. Look at accusations about Google avoiding European taxes by reporting revenue almost entirely in Ireland; Amazon using nonpresence to avoid collecting sales tax as a way of making its products cheaper. Super-intermediaries as non-state but super-state actors.
A: Google’s don’t be evil mantra is itself a sin—it’s pride in assuming you’re not vulnerable to human failings.
Fred von Lohmann: Paper makes a lot of factual assertions; be careful as you flesh out. Struggling with why super-intermediary is an important taxonomy. Can think of 100 companies that satisfy at least many of these criteria; if the issue is speech, then why is this taxonomy special? Broadcasters share many of these; why not ask them to disclose what programs they chose not to run? What stories they chose not to cover? Who influences their decisions? Purpose-built to ID big, consumer-facing internet companies. But why them?
A: relying on his own work right now, but does want further answers.
Jonathon Penney / Internet Censorship, Information Conflict, and the Cycle of Global Telecommunications Technologies
Related to Tim Wu’s cycle of cooptation and innovation, but global communication tech tends to go through a period of novelty and consensus (consensus on promoting the tech and getting it recognized in law & policy); then a period of info conflict and control, when states realize its invasive potential and try to rein it in; then innovation/intervention with a new technology. All of this has happened before and all of this will happen again: undersea cables. The people who read messages sent through undersea cables and decided whether to let them through were called censors. Undersea cables could also be severed with the right equipment.
What was the int’l community’s response to network vulnerabilities? Consensus that it should be protected and promoted. 1875 Telegraph Convention finalized in St. Petersburg: idealistic—a universal right to communicate by telegraph. Notice mechanism to protect messages propogated through network. If a country wanted to suppress at a way station, you had to provide notice to the sending country; no exception to notice requirement for national security or cultural reasons. Expressly allowed for/encouraged encryption for states and private entities. 1884 Cables Convention prohibited severing and provided remedies for parties harmed by severing. No provisions for war.
Highly effective at peacetime; fails when war breaks out. Britain controls the network and realizes that it can use that to its national security and commercial advantage. In 1908, a small war amendment—no notice required if “dangerous” to national security. NSA and GCHQ were built as cable surveillance agencies.
Consensus broke down; paralysis on an international level; mostly unsuccessful litigation on cable cutting, though some success shamign states to pay. Fourth Hague Convention: 1908: cable communications between neutrals are “inviolable.”
Other case studies have a similar pattern.