What is the GRANITE Act?

The GRANITE Act is a bill to stop foreign governments from censoring Americans on the Internet.

On August 27th, 2026, a federal GRANITE Act, the “Guaranteeing Rights Against Novel International Tyranny and Extortion Act,” was introduced in the House of Representatives by Congressman Michael Rulli of Ohio. (See coverage in the Times).

Representative Rulli’s GRANITE Act is the first federal foreign censorship shield bill in American history.

The GRANITE Act has already attracted non-partisan support from NetChoice, ITIF, and R Street. Abroad, the Free Speech Union of Australia has also endorsed the GRANITE Act.

The bill text is available in full at congress.gov.

How a “GRANITE Act” works

A GRANITE Act amends U.S. law to make foreign censorship threats against Americans toothless. The threat alone, even a threat that would be very difficult to enforce, is enough to chill speech and restrict the American marketplace of ideas.

Today, foreign governments threaten Americans for their online speech on a daily basis. Those threats must stop.

All GRANITE bills have the following elements:

1. The shield: they confirm, in statute, the existing judge-made position on non-enforcement of foreign censorship orders, demands, and judgments under First Amendment doctrine;
2. The sword: if a federal bill, they create a private right of action, of some type, that can be brought against a foreign censor (this requirement is optional for state bills, as federal supremacy issues that crop up here have proven tricky in the past); and
3. The definitions: they define “foreign censorship law” and the type of foreign censorship action which engages the shield and, where applicable, the sword in such a way as to address and deter the actual censorship modalities employed by foreign censorship agencies, which are frequently conducted by administrative fiat rather than by judicially-supervised due process.

“GRANITE Act” describes not just Rep. Rulli’s bill, but also other legislative proposals incorporating GRANITE’s three elements

I have been writing about GRANITE for ten months and often refer to the “GRANITE Act” in talks, articles, and other contexts. The term “GRANITE,” as most often used by Internet free speech types when discussing the possibility of a U.S. foreign censorship shield, does not usually refer to a specific bill.

Most of the time, the term describes a bill which implements a particular blueprint for the use of American sovereign power to blunt the effectiveness of censorship orders sent into the United States.

That blueprint is the three-element formula – shield, sword, and definitions – listed above. That blueprint was derived from a decade of experience, on the ground, representing clients, learning how foreign censors seek to project their power into our country.

The GRANITE blueprint has been adapted and modified by different drafters in different ways. There are at least five different public draft versions of GRANITE in circulation today:

• The Wyoming GRANITE Act 2026, HB 70, the first-ever GRANITE bill, which I co-authored, a sword & shield state bill with extensive federal savings language, introduced in February 2026. It passed the Wyoming House of Representatives 46-12 later that month before stalling in the Senate in the short budget session;
• The Wyoming GRANITE Act (2027 reintroduction), a shield-only state bill currently in interim study anticipated to be introduced as a committee bill in early 2027;
• The West Virginia GRANITE Act, a sword & shield state bill, introduced in the West Virginia Senate mere days after Wyoming’s bill, in the short 2026 WV session but did not advance;
• The ALEC GRANITE Act, a model bill spearheaded by Rep. Daniel Singh adopted by ALEC as a model national policy in July 2026; and
• The Rulli GRANITE Act, a federal sword & shield bill, the first federal GRANITE bill.

This list will grow. A New Hampshire GRANITE Act following the ALEC model policy is to be filed in the first or second week of September 2026, with a view towards consideration in the 2027 general session.

I am aware of multiple other GRANITE proposals still in the drafting stages, which I decline to identify at this juncture.

Any bill satisfying GRANITE’s core criteria, as all of the above proposals do, including Representative Rulli’s proposal, would be correctly described as implementing the GRANITE blueprint.

Any federal bill implementing the GRANITE blueprint would, if enacted, represent a dramatic improvement on the present U.S. legal position, where foreign censors avail themselves of the protections of the Foreign Sovereign Immunities Act whilst trampling on Americans’ free speech rights from afar.

I support Rep. Rulli’s GRANITE Act without reservation. Speaking as a practitioner in the (pro bono) business of fighting foreign censors, this bill gives American lawyers the tools that we need to win.

I would support any bill, state or federal, that satisfies the three elements of GRANITE.

1. The Brazen Foreign Censorship Attempt That Led to the Creation of the GRANITE Concept

Readers of this blog will be familiar with the GRANITE concept, which had its origins right here after the UK’s global Internet censorship bureau, the Office of Communications or “Ofcom,” purported to fine my client, the U.S.-based imageboard 4chan, £20,000 for refusing to draft a “risk assessment” – read, a coerced confession – explaining how it, an American business, was guilty of violating UK censorship laws and failing to provide a plan to Ofcom as to how it planned to come into compliance. 4chan, in the lawful exercise of its First, Fourth, and Fifth Amendment rights, refused to comply and sued Ofcom instead in D.C. federal court.

Six months later, Ofcom would purport to fine 4chan a further £520,000 for (a) refusing to dox its users, who have a First Amendment right to anonymous speech and (b) refusing to pay that initial fine.

This purported fine of 4chan was the first time that the UK had ever used the enforcement provisions of its new Internet censorship law, the Online Safety Act 2023, to try to censor Americans. 4chan is by no means the only target: Ofcom has sent at least 197 such notices into the United States between February 2025 and April 2026, only a handful of which are publicly known. Most companies rolled over without a fight: by my count, exactly 4 demands, no more, no less, were publicly refused by Ofcom’s American targets.

Ofcom’s demands are highly intimidating. I have seen several of these notices; each threatened Americans with ruinous fines, and possible prison terms, for non-compliance. This is not an exaggeration: see Ofcom’s correspondence to 4chan, in which the UK threatened exactly that, in writing. The letters were, as many of the letters I have seen from foreign censors are, labeled “CONFIDENTIAL” in shouty all-capital letters. In the lawful exercise of my client’s First Amendment rights, we published the letters anyway on October 16th, 2025.

It appears that most American recipients of these orders, which if recognized in the United States would violate Americans’ First Amendment rights, quietly fold to this foreign pressure rather than seeking to assert their rights.

The UK fined 4chan because the UK thinks that being American on the Internet should be illegal, and that an American exercising his constitutional rights to host content under American law, from an American webserver, in America is subject to the UK’s extraterritorial jurisdiction.

In response to 4chan’s lawsuit, and in correspondence with my client, Ofcom took the position – in writing – that British censorship law overrides the First Amendment on American soil.

Ofcom also asserted that American law on sovereign immunity, specifically the U.S. Foreign Sovereign Immunities Act, shielded Ofcom from being held accountable for its censorship of Americans in an American court.

Put another way, Ofcom wants to have its cake and eat it. The UK’s official Internet censor wants to control American conduct on American soil, and it wants free rein to enforce British censorship law against American citizens, while also being completely unaccountable for its actions in American courts.

That Ofcom, or any other foreign government, can make this argument and not immediately be laughed out of an American courtroom is due to a structural flaw in American law – a flaw only Congress can fix.

2. Sovereign Immunity: the Structural Problem with American Law Where Foreign Censorship is Concerned

Ofcom can take this position due to an unfixed structural flaw in the American legal system, one which arises out of the fact that our sovereign immunity doctrines were designed for an analog world to prevent such absurdities as warships being seized by bailiffs when docked in foreign ports. Our sovereign immunity doctrines were not designed for the situation where a foreign regulator e-mails censorship demands into the United States by the hundreds, then claims the sovereign’s ancient dignities to immunize itself when the American targets of those demands seek redress before an American judge.

The doctrine begins with The Exchange v. McFaddon, 11 U.S. 116 (1812), in which Chief Justice Marshall held that U.S. courts could not be used to seize a French warship docked in Philadelphia’s harbor while sheltering from bad weather in the North Atlantic: the “perfect equality and absolute independence of sovereigns” required judicial hands off. Rules like this made sense when sovereign power projected itself physically but lawfully into the United States, with the nation’s consent, by a hull in a harbor, or an ambassador in a capital city.

Congress comprehensively modernized American sovereign immunity doctrine exactly once, in 1976, before the Internet existed, when the Foreign Sovereign Immunities Act, or FSIA, codified the so-called “restrictive theory” of sovereign immunity which holds that foreign sovereigns and their instrumentalities are, generally, not subject to the jurisdiction of the courts of the United States, subject to certain, narrow, exceptions.

But the FSIA’s exceptions map onto the analog world of commerce, expropriated property, and torts committed entirely in the United States – put another way, the physical world. In many cases, sovereign coercion transmitted by wire fits none of these situations. In Doe v. Federal Democratic Republic of Ethiopia, 851 F.3d 7 (D.C. Cir. 2017), Ethiopia infected a Maryland man’s home computer with FinSpy spyware and recorded his family’s Skype calls; Ethiopia was held immune, because the tort was not committed “entirely” within the United States. In Broidy Capital Management v. Qatar, 982 F.3d 582 (9th Cir. 2020), Qatar hacked an American’s servers and leaked his e-mails to the press; Qatar, too, was found immune, for the same reasons.

Under current doctrine, a foreign state that sells widgets to Americans or employs Americans can be sued. A foreign state that uses the Internet to surveil, hack, or threaten Americans with prison time to intimidate them in the free exercise of their constitutional rights, on the other hand, in each case targeting American citizens on American soil, retains, under our law, for its conduct, the dignified immunity of a foreign warship calmly undisturbed at anchor.

It is time to modernize sovereign immunity doctrine once again. Our law in this area was made for the analog world of ports and paper contracts, not the world we live in now, a world of decentralized packet-switched networks and wires that carry a foreign censor’s threats directly into American living rooms and office suites. Reform is long overdue, to prevent foreign interference with American citizens, exercising American civil rights, on American soil, and online.

For this reason, on October 18^th, 2025, I published v0.1 of the GRANITE Act – a state-level shield bill designed for the State of New Hampshire that touched on all of these points, a practitioner’s wish-list of sorts for the tools that we need to stop foreign censorship cold at the American shoreline, in the hopes that this would serve as a roadmap for Congress to implement a federal fix.

Others quickly followed with their own interpretations. Wyoming Deputy Secretary of State Colin Crossman, Wyoming State Representative Daniel Singh, and I subsequently developed an introduction-ready form of GRANITE, the Wyoming GRANITE Act. Wyoming GRANITE was introduced by Rep. Singh as HB 70 in January 2026 and passed the Wyoming House 46–12 on February 23^rd, 2026 before falling short in the Senate during the short budget session. (Wyoming will attempt to pass GRANITE again, in shield-only form, in 2027.)

Just last month, on July 24^th, 2026, ALEC adopted the “shield” provisions of GRANITE as national model policy, which will accelerate state adoption of shield laws regardless of whether a federal solution materializes. ALEC describes its particular version of the GRANITE concept as “[helping] states shield Americans from foreign laws that attempt to compel speech or incentivize the removal of online content protected by the First Amendment.” Elected representatives in New Hampshire have already pledged to introduce the ALEC bill in the 2027 session.

Of course, a federal solution has materialized: Representative Rulli’s federal GRANITE Act.

3. What the Rulli GRANITE Act Does

Rulli’s federal GRANITE Act shields Americans from foreign censorship. It does so by adding a new section 4106 to title 28 of the United States Code (new 28 U.S.C. § 4106), and by making consequential amendments to the Foreign Sovereign Immunities Act (principally by adding new 28 U.S.C. § 1605(a)(7)).

Chapter 181 is where the SPEECH Act, Congress’s 2010 answer to foreign libel judgments, lives (§§ 4101-4105). GRANITE is to be the SPEECH Act’s neighbor, both conceptually and in the U.S. Code itself.

The bill has four operative parts:

1. a shield (new § 4106(a));
2. a sword/private right of action (new § 4106(b));
3. definitions of “foreign censorship law” and “foreign state” (new § 4106(c)); and
4. a new exception to foreign sovereign immunity (new § 1605(a)(7)).

The Shield – new 28 U.S.C. § 4106(a)

Section 4106(a) contains a statutory shield against foreign censorship judgments, fines, orders, or demands:

Notwithstanding any other provision of Federal or State law, a domestic court shall not recognize or enforce a foreign judgment, or a fine, order, or demand issued by a foreign court or a foreign state, enforcing a foreign censorship law.

Key takeaways:

• The shield applies by its terms to a foreign judgment, or a fine, order, or demand issued by a foreign court or a foreign state.
• “Foreign state” bears its FSIA meaning, 28 U.S.C. § 1603, which sweeps in political subdivisions and agencies or instrumentalities of a foreign state. The shield accordingly applies to censorship orders whether they issue from foreign governments themselves (ministries and law-enforcement bodies, for example), from foreign courts (whether via private right of action or otherwise), or from non-judicial regulatory agencies structured as statutory corporations, such as the UK’s Ofcom.
• Addressing both foreign courts and foreign states is essential to capture foreign censorship as it is actually practiced. Unlike the SPEECH Act, which dealt only with foreign judgments, foreign censorship orders are frequently issued by foreign states directly without any judicial supervision whatsoever. UK Ofcom, for example, is a “foreign state” for the purposes of the FSIA and purports to issue binding orders and demands on Americans without judicial supervision; other foreign states attempt to censor Americans with their police forces, such as the German BfJ in its handful of attempted extraterritorial NetzDG enforcement actions and French prosecutors seeking to compel information from Elon Musk and Linda Yaccarino (discussed below).

Addressing all foreign judgments, and not only orders issued by foreign censorship agencies, is important if the GRANITE Act or similar proposal is to be future-proofed. British censorship advocates, such as Baroness Beeban Kidron, have proposed that, given the failure of agencies like Ofcom in the handful of cases where Americans have publicly refused Ofcom’s orders (note: I am defense counsel in all of those cases), future UK censorship statutes should deputize British tort claimants to carry out censorship-by-proxy of American citizens, as reported in the Observer in July.

This censorship modality – deputizing private actors to do what the state itself cannot – would be covered by Rulli’s GRANITE shield language also, thus ensuring American judicial machinery cannot be used to enforce a foreign censorship order no matter what form that judgment takes. I add that, although there is an argument that the “use torts where Ofcom has failed” approach is foreclosed by 28 U.S.C. §4104 of the SPEECH Act, §4104(b) makes it plain that the application of §4104 is intended for defamation judgments only. To address the wider range of foreign censorship we see today, a new statute is needed.

GRANITE’s language would therefore also shield Americans from other, new areas of foreign content regulation, not strictly related to online content regulation or political censorship, but which nevertheless have the substantial effect of censoring online speech in the United States, in the rapidly-expanding censorship regimes of America’s allies. These proposals include, e.g., the UK Law Commission’s law reform proposal which would create civil and criminal liability for both speakers and platforms for reporting on ongoing court proceedings, or hosting that speech – acts which are, quite unambiguously, constitutionally protected in the United States or on American platforms.

The statutory shield should, on plain and ordinary construction, also be usable by secondary targets of business disruption orders, such as search engines or VPN/CDN providers like Cloudflare, to seek relief in the event that foreign censors seek to serve business disruption orders and the like on these companies in order to affect other American businesses who are the primary target of the order.

The Sword – new 28 U.S.C. § 4106(b)

Second, federal GRANITE creates a private right of action against a foreign government or sovereign instrumentality, like Ofcom.

The purpose of the “sword” is to deter foreign censorship threats from entering the United States, keeping in mind the 197-notices-and-4-refusals number mentioned above which shows that most American companies, when threatened by a foreign censor, surrender without a fight.

This is a cousin of the Brussels Effect, the “London Effect” if you will, except instead of market access driving compliance, foreign censorship is driven by threats and intimidation. Furthermore, the First Amendment, not user data or business combinations, is in foreign censors’ crosshairs. In the absence of any American sovereign response, the UK looks like it’s getting away with it, which is why everyone else is trying to get away with it too – Brazil, South Korea, Australia, and the EU included.

American free speech doctrine has long recognized that threats, even if those threats are difficult to enforce, chill speech. Foreign censors’ threats chill American speech. Accordingly, the threats must stop, and the only way they will stop is if the act of threatening an American for constitutionally protected activity online leads to unpleasant consequences for the foreign entity making that threat.

The “sword” language reads:

A United States person against whom a foreign judgment described in subsection (a) is entered, or upon whom a judgment, fine, order, or demand described in subsection (a) is served or purportedly served, may bring an action in an appropriate district court of the United States seeking declaratory relief that such judgment, fine, order, or demand is unenforceable under subsection (a).

Breaking that down a bit:

The wrongful act that GRANITE restrains with the private action is (i) enforcing a foreign censorship judgment against an American, or (ii) serving or purporting to serve a judgment, fine, order, or demand on an American that, if recognized in America, would violate the First Amendment.

“Fines, orders, or demands” are referenced here because most foreign censorship regimes are extrajudicial in character. The UK’s Ofcom, for example, purports to be an “independent regulator” and not formally part of the government at all in the UK’s domestic legal environment (despite claiming to be a sovereign instrumentality for U.S. law purposes); the German Federal Office of Justice (BfJ) purports to be able serve binding orders across America’s borders despite being formally part of the German government.

In most cases, these censorship orders are purportedly legally binding, would violate the U.S. Constitution if recognized here, and are not subject to judicial supervision at all.

I want to draw attention to the language “purportedly served.” Ofcom’s practice, documented across aforementioned 197 separate notices sent into the United States as of April 2026, is to e-mail demands and orders across the Atlantic in a manner that, whilst authorized and indeed required by the Online Safety Act, would get a junior law firm associate fired if attempted in civil proceedings.

Normally, service of legally binding process across an international boundary requires adherence to a treaty (an MLAT, for example) or an international service convention like the Hague Service Convention. Foreign censors know, however, that American courts are unlikely to accord reciprocity to their orders if they seek formal recognition for those orders here. The censorship codes they enact overseas therefore purport to authorize regulators like Ofcom to serve Americans with legally binding orders by e-mail, which U.S. law does not recognize. The result is foreign censorship agencies threatening and bullying Americans by e-mail.

Under § 4106(b), the defective service with which Ofcom purports to bind American citizens is itself sufficient to bring an action against Ofcom, or against any other foreign censor who uses the same methods, as all of the major western Internet censorship states do.

This would also permit Americans who receive foreign censorship orders in relation to which they are secondary, but not primary targets of the censorship in question, such as Cloudflare or payment rails, to refuse the order and sue. We have already seen the UK advocating for expanded use of so-called “business disruption orders” against American companies, in order to try to starve American companies off the Internet by cutting them off from essential counterparties where they cannot force them off of the Internet. The American censorship deterrent needs to address that censorship modality, and this language does.

The private right of action should, on plain and ordinary construction, also be usable by secondary targets of business disruption orders, in much the same way as the shield component is.

The relief is a declaratory judgment.

Definitions – new § 4106(c)

“Foreign censorship law” is defined at § 4106(c)(1):

any law of a foreign state that— (A)(i) has the primary purpose or substantial effect of restricting or penalizing expression, expressive conduct, or expressive association; (ii) targets expression, expressive conduct, or expressive association based on its content, viewpoint, or speaker identity; or (iii) compels disclosure that would reasonably chill that expression, expressive conduct, or expressive association; and (B) would conflict with the protections of the First Amendment of the United States Constitution if recognized in the United States.

“Foreign state” is defined at § 4106(c)(2) by reference to the definition of “foreign state” in the FSIA, which also includes sovereign instrumentalities.

A foreign law falls within the definition of “foreign censorship law” if it does any one of the three things listed, and would conflict with the First Amendment if recognized here. Accordingly, this language should not affect legitimate cross-border law enforcement cooperation between the United States and other nations for online crime, as “foreign censorship laws” are only those which have the primary purpose or substantial effect of penalizing First Amendment protected speech, target protected speech, or seek to compel disclosure that would chill protected speech.

“Expression, expressive conduct, or expressive association” is a broad sweep, and would include not only political speech but also software development, including non-custodial DeFi software publication and AI model development, and web hosting. This language should shield not only social media companies, but also advanced technology companies, including AI developers, from foreign regimes, ensuring that U.S. law can be selected as the governing law for U.S. technology firms.

The language “primary purpose or substantial effect” would capture most foreign censorship statutes incompatible with the U.S. Constitution, whether they are intentionally censorial (such as Section 127 of the UK’s Communications Act 2003, which criminalizes “grossly offensive” speech) or simply incidentally censorial as-applied (such as Section 5 of the UK’s Public Order Act 1986, which criminalizes speech based on its subjective perception by a hearer as causing alarm or distress).

The bar on compelled disclosure would also protect Americans from intrusive investigative demands which frequently precede formal enforcement action, and which are directed towards the censorship of protected speech online. These demands should be properly routed through the applicable MLAT procedure and subject to American judicial supervision. Foreign censorship statutes have been written to circumvent that procedure on account of the fact that their drafters knew that foreign censorship orders could not survive American judicial scrutiny.

The definition would capture censorship that foreign states enforce directly, through sovereign organs and their instrumentalities, and censorship they enforce indirectly, for example, through the creation of novel private rights of action in their own courts. Whether to make this more explicit is a question for Congress. The use of private rights of action as a proxy for sovereign censorship currently exists under the GDPR and has been proposed as a possible future extension to the UK Online Safety Act, as mentioned above.

Amendment to the Foreign Sovereign Immunities Act – 28 U.S.C. § 1605(a)(7)

This is the shortest provision in the bill, but also the most important, for it amends § 1605(a) of the Foreign Sovereign Immunities Act to add a new exception to foreign sovereign immunity for cases arising under § 4106 under new § 1605(a)(7).

That new paragraph (7) is exactly three words long:

“under section 4106.”

For its brevity, it does a great deal of work: without it, a foreign censor would answer a § 4106(b) lawsuit by invoking sovereign immunity and moving to dismiss under FRCP Rule 12(b)(1). In most cases, at that point, the case would end. That is one of the defenses Ofcom has raised in 4chan’s D.C. litigation, and one that Brazil has raised in Rumble Inc. and Trump Media and Technology Group Corp. v. De Moraes.

With this exception on the books, however, foreign censors could be sued for attempts to censor Americans in American courts.

4. The Rulli GRANITE Act in the Broader Context of Current U.S. GRANITE Proposals

Legislation is as critical a battleground as First Amendment litigation for free speech on the Internet, given the international relations dimension to the problem

4chan’s fight with the UK’s Ofcom is not the point of the GRANITE story. The 4chan case was simply the first example, the UK’s first-ever attempt to extraterritorially enforce its online censorship law against an American target.

From the UK’s perspective, the 4chan case has been an unmitigated public relations disaster.

From the American perspective, the first-ever enforcement could not be allowed to succeed, because success in that action would set the precedent for every action that followed it. 4chan’s legal team is determined to ensure that our client, not Ofcom, will be the precedent-setter in this dispute. We don’t want what happened to 4chan to happen to anyone else in the United States. The GRANITE blueprint was therefore proposed to protect all Americans from the foreign censorship threat.

From the outset, it was clear that, whatever its outcome, the 4chan dispute would become a case study. 4chan did what any American exercising constitutional rights is entitled to do: it publicly refused Ofcom’s orders, invoked the First, Fourth, and Fifth Amendments, and told the UK to come back with a warrant. Ofcom, for reasons grounded in the UK’s domestic political environment, responded to this by

• ignoring every legal cooperation treaty between the UK and the United States,
• purporting to serve 4chan with process binding in the United States by mere e-mail,
• fining 4chan, whose operations are based exclusively in the United States, close to $800,000;
• threatening 4chan’s operators and volunteer administrators with arrest;
• asserting, in writing, that the UK had the power to override the First Amendment, and
• invoking sovereign immunity in an American court when challenged,

…all in response to 4chan’s conduct which is perfectly lawful, and indeed constitutionally protected, in the United States.

Congress has noticed.

Ofcom’s conduct in this case is a preview, a demonstration of what will happen to every other American Internet company, hundreds of thousands of them, and potentially millions of American citizens if Congress does not intervene to halt foreign countries’ attempts to censor Americans from afar.

The next frontier of the fight for free speech is transnational, because most speech happens online and the Internet is, fundamentally, global. The clash over that domain is being, and will continue to be, waged by sovereigns, and the great global victories and defeats for free speech on that global system of communication will not be won or lost in courtrooms, but in legislation and treaties that expand or constrain sovereign power.

This differs from traditional First Amendment fights, the landmark ones fought by Abrams and Bickel in New York Times v. United States, or David Goldberger in National Socialist Party of America v. Village of Skokie, where principled American litigators have traditionally taken on unpopular causes against censorial tendencies in the American government.

In a cross-border context, American courts – without more – can do little. The job of the “free speech lawyer” will become, in part, to enlist the American government in the defense of Americans’ rights, which are now endangered by other states. On the Internet, American domestic speech rules mean nothing unless they are backed by American power, and are drafted so that Americans can impose costs – political and financial – on foreign states for any attempt to interfere with our rights.

On the transnational front, free speech has so far seen an unbroken run of defeats: the EU DSA and AI Act, UK/Australia Online Safety Acts. All constitute extraterritorial power projection into the United States.

Rep. Rulli’s GRANITE Act is designed to counter that power projection and would represent, if enacted, the first time that free speech advocates meaningfully reversed the extraterritorial censorship tide.

Rulli’s GRANITE as a proportionate American response

I should add that Rep. Rulli’s bill is, to my knowledge, not the only GRANITE-type effort underway. Others, in Congress and in the States, are thinking seriously about how best to shield Americans from foreign censorship. Some of these proposals, such as Wyoming’s HB 70, included such provisions as statutory damages, an express bar on executive branch cooperation, and extradition protections that are not explicit in Rep. Rulli’s bill, but arguably not required by it, either. Whether a future proposal chooses to be more explicit on these points is a matter for the proposals’ drafters.

These are all good ideas. Whether they can be enacted in this Congress or a future one is another matter. What Congress ultimately decides to expressly protect, and how, is a political question.

In my view, the damages issue is really the only material difference between Rulli’s bill and a line-for-line federalized version of Wyoming’s HB 70. Bars on extradition and executive-branch cooperation address important but narrower censorship modalities, such as the recent attempted abuse of the MLAT procedure by France to attempt to serve demands on Linda Yaccarino and Elon Musk for protected speech, which remain natural candidates for introduction and/or harmonization in committee markup, although Rulli’s First-Amendment-grounded shield already establishes the core principle in relation to which any additions of this type would be complementary.

Due to the FSIA, for most disputes in American courts on the subject of vindicating an American’s free speech rights online, the deck will be stacked in the foreign censor’s favor. Rep. Rulli’s GRANITE Act flips that situation and will help ensure that every American lawyer will be far better equipped to make any attempt to censor Americans online an extremely unattractive proposition.

The original GRANITE Act design from October 2025, and the version of HB 70 advanced in Wyoming in early 2026, was designed to be a thermonuclear-tier deterrent to foreign states attempting to censor Americans online. Under that proposal, a regulator that threatened an American with a $25 million fine would face a statutory damages floor in a Wyoming courtroom equal to the greater of $1 million or the maximum amount of the threatened fine. As the UK’s demands bear a maximum fine equal to the greater of 10% of worldwide turnover or £18 million (~$25 million), the Wyoming proposal, if federalized, would have the result that Ofcom’s entire annual “Online Safety” budget (£72 million) would potentially be forfeit after sending only four e-mails into the United States.

Of course, HB 70 was a state bill. The damages language was heavily caveated with federal savings language to account for FSIA, Commerce Clause, and foreign relations issues. The idea was that Wyoming could turn itself into something of a porcupine in the event that a federal shield bill failed to materialize; even though FSIA issues are difficult to get over, they are not impossible, and it was, theoretically, possible to create enough doubt in the mind of a foreign censor that the censor would have to think twice before targeting a Wyoming company.

The damages language in the October 2025 proposal served one purpose: to make the act of attempting to censor American citizens so unpleasant for a foreign government that the behavior stops. It was never intended as an invitation to build a thriving plaintiffs’ bar that nitpicks foreign legal orders, finds First Amendment violations, and raids foreign sovereign assets custodied at American banks. It was intended as a deterrent that would force foreign legal demands back into treaty channels, in particular the MLAT channel, where the requests can be supervised by the U.S. Department of Justice and Department of State. At the moment, foreign governments are conspicuously ignoring treaty channels, and the comity principles these channels are designed to protect, by purporting to serve foreign censorship orders on Americans directly by e-mail.

Rulli’s federal bill differs from what was envisioned by HB 70 in that it does not propose to award damages to a successful plaintiff. Rep. Rulli’s bill proposes only to allow U.S. plaintiffs to obtain declaratory relief. In my opinion this likely will make Rulli’s bill easier to pass than a bill with extensive damages language (and thus requiring more invasive surgery on the FSIA), although this is just a hunch.

In that respect, Rep. Rulli’s bill is also different, and more aggressive, than the ALEC model bill referenced above, which is shield-only on account of the fact that sovereign immunity is a federal, rather than state, question.

The structure of Rep. Rulli’s bill provides different disincentives to the censorship of Americans, and my experience with the 4chan case tells me that they are likely to be effective ones.

Shield-as-disincentive

The shield itself is one form of disincentive. Thanks to a freedom of information request about the 4chan case, we now know Ofcom spent between 0.9 and 2.6 full-time-equivalent staff per month, for nine months – assuming an average of 2.0 FTE, ballpark, this works out to something like 2,800 man-hours in all on the 4chan enforcement action, even before considering the cost of related litigation. As is known to those who have been following that case, these thousands of hours of regulatory labor yielded a return of (a) zero compliance, (b) one federal lawsuit, and (c) exactly four emails featuring hamsters that took me, as 4chan’s counsel, perhaps thirty minutes, in total, to produce.

Over the course of the 4chan enforcement, we also observed that British censors and the British political class have, broadly speaking, chosen not to acknowledge that the First Amendment exists. They maintain the fiction – for a domestic UK audience and their foreign targets alike – that UK censorship orders stand some slim chance of being enforced in the United States. When one American target, my client, refused out loud, the UK went so far as to threaten to deputize American law enforcement to collect.

It is, of course, possible that a first-instance U.S. court could get the law badly wrong and accord reciprocity to a foreign censorship order. Ofcom’s bluff, in raw numerical terms, appears to have been a highly effective one, as 193 of their Section 100 orders sent into America were, apparently, obeyed where only 4 – all my clients – were publicly refused.

Rulli’s federal GRANITE Act would foreclose the possibility of any American court erroneously according reciprocity to these orders, would embolden American targets of foreign censorship to refuse the orders, and, in doing so, would force foreign censors to concede, to their funders and their targets, that every penny spent chasing Americans is a penny wasted. Judicial errors of this kind are not hypothetical; in 1995’s Stratton Oakmont v. Prodigy, a first-instance judge in the New York Supreme Court mangled intermediary liability doctrine so badly on a novel Internet question, that it required Congressional intervention, the Cox-Wyden amendment that became 47 U.S.C. § 230, to fix.

A shield statute like Rulli’s GRANITE would, therefore, guarantee that foreign censorship agencies seeking to project their rules into the United States would meet stiffer resistance and have less political air cover at home. It would not reach into foreign countries and affect what domestic laws they can enact for themselves.

Post-Rulli-GRANITE, small to mid-size companies with territorial footprints bounded to the United States would be able to operate from the U.S. safely with absolute confidence.

Post-Rulli-GRANITE, global companies with direct territorial exposure (such as data centers, power generation, bank accounts, or employees) in pro-censorship countries like the UK, Europe, Brazil, China, or elsewhere, that wish to provide uncensored, globally-accessible platforms, will need to weigh whether divesting from those jurisdictions in order to benefit fully from the American shield is a worthwhile tradeoff.

Sword-as-deterrent

The sword in Rep. Rulli’s bill, which does not have damages language, needs less doctrinal machinery than would be required by a proposal that federalized the hefty damages language from Wyoming’s 2026 HB 70. Because the only relief on offer is declaratory, the FSIA exception has been narrowly drafted; unlike the original GRANITE proposal, execution immunity under §§ 1609-1611 is simply not an issue, the attachment and asset-seizure questions that complicate FSIA litigation drop out entirely, global financial rails that run through the United States are not implicated, and the State Department is spared any major diplomatic headaches.

Keeping in mind the fact that I co-authored HB 70, my own preference – no secret to anyone who read that proposal – would be for hefty statutory damages language. I am cognizant, however, that where FSIA doctrine is concerned, incremental changes may be easier to get through Congress than dramatic changes. Rulli’s approach should be, comparatively, much easier to pass than a bill which federalized the approach we took with HB 70.

Rulli’s sword is still an effective one. In terms of satisfying the core GRANITE criteria, it would create a judicial remedy that would allow a foreign censor to be dragged in front of an American judge and held there, for an extended time, involuntarily, for as long as the litigation lasts. It is my impression that foreign censors and their political masters dislike being sued in America intensely, and that having their domestic censorship regimes labeled as such in American court filings is politically damaging for them at home.

GRANITE as a shot across the bow

An e-mailed demand costs a foreign regulator nothing to send today and is frequently cloaked in secrecy. I have lost count of the number of times a foreign censor has put CONFIDENTIAL in all-caps or bright red text, or even threatened a client with jail time (as Australia did to one American client), to prevent publication of correspondence the foreign censor has transmitted into the United States that the American recipient has every right to publish, and that the American public has a right to know.

Post-Rulli-GRANITE, every e-mailed demand of this type into the United States is an invitation to the American target to drag the sender into proceedings in an American forum, which is expensive, time-consuming, embarrassing, and public.

Hopefully, enacting even a restrained amendment to the FSIA – keeping in mind that the FSIA is itself not holy writ, but a gift that the American Congress extends to foreign states in the interests of comity – will demonstrate how seriously the United States takes attempts by foreign governments to censor our people.

Foreign governments are no longer extending comity to us, where our Constitution is concerned, on the Internet. It is only fair that we return the favor.

If, in a post-Rulli-GRANITE world, foreign regulators one day conclude that pursuing Americans is not a waste of money, and that the risk, bad optics, and expense of being dragged into American litigation is an acceptable one, Congress retains every option to increase the costs associated with foreign censorship targeting of American citizens with a simple amendment to new § 4106.

Whether Congress eventually chooses to escalate to a full fine-matching statutory damages regime combined with a novel targeted exception to execution immunity, as Wyoming’s HB 70 suggested be done earlier this year, would be governed entirely by foreign censors’ future behavior.

I strongly suspect, however, that Rulli’s GRANITE Act will prove to be enough to keep the censors out of most Americans’ inboxes for the foreseeable future.

Congress would do well to pass it.