"State Regulation of Online Behavior: The Dormant Commerce Clause and Geolocation"

Prof. Jack Goldsmith (Harvard) and I will have this article out in the Texas Law Review early next year, and I'll be serializing it here in the next couple of weeks, starting Tuesday. There is still plenty of time for editing, so we'd love to hear any recommendations you folks might have; in the meantime, you can read the entire PDF of the latest draft (though with some formatting glitches stemming from the editing process) here. Meantime, the Abstract:

When does the Dormant Commerce Clause preclude states from regulating internet activity—whether through state libel law or invasion of privacy law; through state laws requiring web sites to accommodate disabled users (for instance, by providing closed captioning); through state bans on discriminating based on sexual orientation, religion, or criminal record; or through state laws that ban social media platforms from discriminating based on the viewpoint of users' speech?

This essay argues that the constitutionality of such state regulation should generally turn on the feasibility of geolocation—the extent to which web sites or other internet services can determine, reliably and inexpensively, which states users are coming from, so that the sites can then apply the proper state law to each user (or, if need be, choose not to allow access to users from certain states). In recent years, geolocation has become feasible, and is routinely used by major web sites for ordinary business purposes. There is therefore more constitutional room for state regulation of internet services, including social media platforms, than often believed.

Note that the article is exclusively about the Dormant Commerce Clause questions, not the separate defenses internet services might have under either the First Amendment or 47 U.S.C. § 230 (for more on that, see this article of mine and this article by Adam Candeub and me).

Comments

[Stephen Lathrop]

There is therefore more constitutional room for state regulation of internet services, including social media platforms, than often believed.

But so much of the likely impetus to regulate social media platforms will come from unconstitutional attempts to violate the 1A press freedom rights of publishers.

[Bored Lawyer]

But that's not his issue. The Dormant Commerce Clause is a separate doctrine that the states cannot interfere with interstate commerce. Congress, on the other hand, can.
The First Amendment, OTOH, limits both the federal and state governments.

[Michael P]

And by "1A press freedom rights of publishers" should we assume you mean "statutory privileges for interactive computer service providers that do not obtain for publishers more generally" (especially when they censor viewpoints in a partisan manner)?

[Stephen Lathrop]

Michael P, no statutes needed. All publishers, online or otherwise, enjoy 1A liberty to publish at pleasure and, with the exception of libel law, without constraint. They can publish whatever they please and reject the rest, without regard to viewpoints or anything else. You are at liberty to set up as a publisher and do the same.

That is such a firmly established constitutional protection that advocates of all kinds of viewpoints have been trying furiously to figure out some way to make the law call online publishers by some other name, on exactly the same deluded principle as Lincoln's 5-legged dog. If those advocates ever get their way, that will be the end of press freedom in the U.S.

[Michael P]

Really? You think publishers can publish without regard to laws about child pornography, copyright infringement, extortion, and so forth? How has that worked out for Mindgeek?

You shouldn't assert such falsehoods so confidently.

[Bored Lawyer]

You are avoiding his question, though. Libel law generally makes publishers liable for third-party libel they publish. On-line publishers are exempt from that by federal law.

Personally, I would not abolish that, but I would make it contingent on an honest and full disclosure of what basis the platform censors content on its platform. Deviate from your full disclosure, you lose your immunity.

[Kazinski]

Unfortunately the first amendment only applies to states and Federal government, but it doesn't apply to private companies or foreign governments.

And as Stewart Baker often points out Europe and China care not a wit for our freedom of speech sensibilities.

[David Nieporent]

More accurately, the first amendment does apply to private companies, in that it protects them from government meddling into what speech they distribute.

[Eugene Volokh]

Still more accurately, the First Amendment protects private companies from government meddling into what speech they distribute in some situations, see, e.g., Miami Herald; Hurley; PG&E, but not others, see, e.g., PruneYard; Turner; FAIR (and this article).

[David Nieporent]

And by "1A press freedom rights of publishers" should we assume you mean "statutory privileges for interactive computer service providers that do not obtain for publishers more generally"

Even if you didn't know Lathrop's pathological hatred of free speech for everyone but newspaper publishers, why on earth would you "assume" that he meant the exact opposite of what he said?

[Michael P]

Because he regularly conflates the First Amendment with Section 230.

[Stephen Lathrop]

?

[Michael P]

They can publish whatever they please and reject the rest, without regard to viewpoints or anything else.

That's not true in the general case; all the kind of laws I mentioned above limit a publisher's discretion, and I only listed the most significant/common categories. For example, New York Times v. Sullivan didn't hinge on the NYT being the publisher of the advertisement, but whether the First Amendment protected the advertisement at all. The only time a publisher as much discretion as you suggested is when they're protected by Section 230.

[Stephen Lathrop]

Nieporent, publishing generally—not just newspaper publishing—has been the greatest force multiplier ever created for speech freedom. My advocacy on behalf of press freedom is made with that always in mind.

Separately, I am also an advocate of private editing prior to publication. You differ on that, and seem to oppose private editing as oppressive. I think that advocacy is utopian. I insist press freedom will not long be tolerated politically without some means to keep libel in check.

I think you already know all that. My comment is directed at other readers who might not know what to think about terms such as, "for everyone but newspaper publishers," and "pathological hatred," when you use them.

[David Nieporent]

Separately, I am also an advocate of private editing prior to publication.

Once again: you are an advocate of government coercion to force private editing.

My comment is directed at other readers who might not know what to think about terms such as, "for everyone but newspaper publishers," and "pathological hatred," when you use them.

You might want to disclose to said "other readers" that you also oppose fair use, student free speech, politician free speech, anti-vaccine speech, etc.

[Rossami]

I'm not yet sure where you're going with the article but I think you might have a basis premise wrong. Geolocation is not now, never has been and never will be sufficiently reliable to be a basis for legal decisions. There are far too many confounding factors, both intentional and unintentional.

Consider first that most geolocation services report the IP location not of the individual user but of their ISP (or perhaps an ISP substation). Granted, for most people who live in the middle of their state, you'll get the same answer but for the many people who live near the border, whether the state can be correctly identified becomes near-random.

Consider next that many people use VPNs. (Recent data suggests that when you count corporate users, the percentage is a clear majority of connections using VPN.) VPN makes location data unreliable on purpose.

If you limit your analysis to mobile devices (phones and ipads) that have GPS, remember that a growing population are learning to turn them off by default. Consider also that many companies don't want to know your precise location (which, again, becomes necessary if you're near a border) because of the increased regulatory burdens that attach if you collect "precise" geolocation data.

[Rossami]

Argh, typo. Should be "basic premise" in the first row.

[Kazinski]

That's true, but I think the workaround is a popup that shows your presumed location, then you have to certify that's your location, or you are an unauthorized user.

Hopefully they can't prosecute or sue companies that are victims of location fraud.

I've got location sharing off, I'm not naive its foolproof, but it works at least a little, Home Depot thinks I'm in Sacramento, Lowes thinks I'm in Seattle, neither one is close.

[Eugene Volokh]

Rossami: Here's what we say in the article (and what we'll quote in a later post in the series):

Geographical identification and filtering on the internet have improved enormously in the last two decades, but they remain imperfect, and there is a persistent arms race at the margins between blocking and evasion technologies. Still, geolocation technology doesn’t have to be perfect to be useful for legal compliance, just as many laws and security technologies remain useful even if they are not perfectly enforced. That would be especially true if a state law treats reasonable use of geolocation technology as an adequate defense when geography matters—for instance, when reasonable even if imperfect geoblocking attempts are seen as evidence of lack of intent to target a particular state for personal jurisdiction purposes, or when reasonable geolocation attempts are seen as sufficient for determining that a defendant’s copyright infringement happened within the court’s jurisdiction....

The pervasive reality of geolocation and filtering services on the internet should affect Dormant Commerce Clause analyses of state internet regulation. Most courts that examine state internet regulations do so under the undue burden test and some version of the extraterritoriality test. Whichever of these tests courts deploy, and whatever version of the extraterritoriality test they use, the analysis in Parts II and III suggests that four principles related to the geographical element of online transactions should govern the analysis.

First, the internet is not a borderless medium. All major firms operating on the internet, and many smaller ones, collect and use location data about consumers and users, and shape content by geography. The technology that supports these practices is quickly growing more pervasive, more accurate, and less expensive.

Second, because geolocation and filtering technology is pervasive, courts should not presume that internet operators have any greater difficulties than “real-space” operators in identifying internet users based on geography and tailoring their products to state law. Assessing the costs and benefits of complying with state regulations, or of the extraterritorial impact of state regulations, must include realistically assessing compliance costs based on the current state of geolocation and filtering technology.

In this regard, courts should consider the ways that the firm challenging state regulation on Dormant Commerce Clause grounds uses geographical-identification technology to further its business interests. They should also consider the extent to which the costs of such technology can be decreased by the legal regime. For example, as courts have made geo-identification and filtering technology more relevant to personal jurisdiction, firms have increasingly deployed such technology to avoid activities that may expose them to personal jurisdiction, which has contributed to the expanded market for and lowered costs of such technology. Similarly, a Dormant Commerce Clause jurisprudence that accommodates state differences will contribute to more sophisticated and less expensive geofiltering tools....

Fourth, an online firm’s inability to perfectly comply with a state regulation due to imperfections in the accuracy of geographical identification and filtering technology does not by itself mean the state regulation violates the Dormant Commerce Clause. The relevance of technological imperfection will depend in part on how demanding the state regulation is. Several state laws regulating minors’ access to pornography online have survived Dormant Commerce Clause scrutiny because they criminalize only intentional transmissions of banned materials to minors or to minors within a state. For similar reasons, Dormant Commerce Clause concerns are significantly reduced if state law provides a defense for reasonable efforts to keep forbidden internet content out of the state. And the Supreme Court has stated that a State’s efforts to minimize the interstate impact of a regulation, including through compliance software, is relevant to the discrimination and undue burden analyses.

[John F. Carr]

Geolocation is very good in my experience, living in a densely connected part of Massachusetts. My wired address appears to be in the next town a couple miles away, which is probably accurate if you trace the fiber coming out of my basement. When a page for a big chain chooses my local store it's usually right.

[Eric Owens]

It was already ashitty deal that streaming content providers choke on trying to stay internet safe with a vpn.

[Life of Brian]

streaming content providers choke on trying to stay internet safe with a vpn

"Choke" coincidentally describing the effect on your VPN provider when a sizable number of subscribers try to stream at the same time.

Joking aside, I'm unaware of any cogent security issues posed by streaming that a VPN could address. VPN use for streaming is purely to bust geographic restrictions. Whether and under what circumstances that's a just cause feels like a different issue.

[ReaderY]

Professor Volokh,

Perhaps, in introducing the topic, you might want to explain the Dormant Congress Clause argument that your paper explains is less an issue than prrceived. What is perceived here? In what ways is it argued that the Dormant Commerce Clause inhibits stage regulation of the internet? Is the argument that the internet is always interstate commerce?

[Michael P]

I infer that the argument is that when a state's regulations affect traffic destined to residents of multiple states, that is more of a regulation of interstate commerce than when the traffic itself crosses state lines. (Nowadays, the ubiquity of content distribution networks means that much traffic to an end user can reach that user without crossing state lines.)

[Michael P]

Left wing activists also claim a lot of other stupid things. It's a reflection of their mental processes, not the rest of the world.

[Michael P]

It shows that some laws can reach publishers, even if they're not libeling anyone. I'm a little surprised that section 230 hasn't protected Mindgeek, but they're unwoke, so they don't get the full set of privileges.

[Michael P]

Every accusation you make is a confession, as you often advertise.