Censorship by Proxy: When Government Doesn't Need to Ban the Speech — Is America Better Yet?
By Roberto Zapatero
Independent Forensic Investigative Journalist
Is America Better Yet?
Living Record — August 18, 2026
A government does not always have to ban speech to make it disappear.
Sometimes it only has to reach the company that controls whether people can see it, download it, publish it or distribute it.
That distinction is becoming increasingly important in the United States.
In October 2025, the U.S. Department of Justice contacted Apple about ICEBlock, an application that allowed users to report publicly observable activity by Immigration and Customs Enforcement (ICE).
Apple removed the application.
Then-Attorney General Pamela Bondi publicly said the Justice Department had demanded its removal.
That does not, by itself, prove that the United States has become an authoritarian state.
It does raise a question that a democracy cannot afford to ignore:
Can government officials accomplish through private companies what the Constitution may prevent government from doing directly?
What happened to ICEBlock
ICEBlock allowed users to share reports about ICE activity in their communities.
Federal officials sharply criticized the application and its developer, Joshua Aaron.
By October, the Justice Department had contacted Apple and demanded that the application be removed.
Apple complied.
Aaron and his company later sued Bondi and other federal officials. Their lawsuit alleges that the government retaliated against Aaron and unlawfully coerced Apple into suppressing constitutionally protected speech.
That case remains important because it puts the government's conduct directly before a federal court.
But it is no longer the only court case testing this kind of pressure.
A federal judge has already intervened
In a separate case, Rosado v. Bondi, Kassandra Rosado and Kreisau Group LLC challenged federal pressure involving another ICE-tracking application, Eyes Up, and a Facebook group called ICE Sightings - Chicagoland.
On April 17, 2026, U.S. District Judge Jorge L. Alonso granted a preliminary injunction.
The ruling was not a final judgment after trial. But the judge found that the plaintiffs were likely to succeed on their claim that federal officials violated the First Amendment by coercing Facebook and Apple.
The court found that Bondi and then-Homeland Security Secretary Kristi Noem had demanded, rather than merely requested, action from the companies and had made statements the court viewed as intimating the possibility of prosecution.
That distinction matters.
The case provides judicial evidence that the constitutional concern is not merely hypothetical.
The constitutional line
Government officials are allowed to criticize speech.
They can condemn it.
They can argue against it.
They can try to persuade private companies.
But there is a constitutional boundary.
In National Rifle Association of America v. Vullo, the Supreme Court of the United States unanimously reaffirmed in 2024 that government officials cannot use the power of their offices to coerce private parties into punishing or suppressing disfavored expression.
The First Amendment problem does not disappear merely because a private business ultimately carries out the government's wishes.
That creates a critical distinction:
Persuasion is permitted. Coercion is not.
The difficult question is determining where one ends and the other begins.
In Rosado, a federal judge concluded that the plaintiffs had shown enough evidence of coercion to justify preliminary court intervention.
Government does not own the gate
The modern communications system has created something earlier censorship battles rarely confronted at this scale.
Private companies operate enormous gateways to public communication.
Apple's App Store is a central distribution point for applications used on iPhones.
Google operates Google Play and much of the infrastructure surrounding Android application distribution.
Facebook and other social-media platforms provide channels through which millions of people communicate and exchange information.
These businesses are private actors.
The First Amendment generally restricts government, not private companies acting on their own.
But the constitutional question changes when government officials use public authority to pressure a private company into suppressing speech.
The company may push the button.
The critical question becomes: Whose power caused it to be pushed?
A censorship infrastructure does not have to begin as censorship
The Electronic Frontier Foundation (EFF) raised a related concern in November 2025 about Google's plans to verify Android developers.
Google says developer verification is a security measure intended to make it harder for malicious developers to distribute harmful applications anonymously.
That concern is legitimate.
Malware is real.
Fraud is real.
Applications can steal information, compromise devices and harm users.
But infrastructure created for a legitimate purpose can acquire another use.
A centralized system capable of identifying developers and controlling which applications can be installed can also become a point of government pressure.
That does not establish that Google's system was created for censorship.
Is America Better Yet? (IABY) found no public evidence in the sources reviewed establishing that the Trump administration ordered Google to create its developer-verification system for political purposes.
And Google's program has changed since EFF published its warning.
Google now says students and hobbyists can use limited-distribution accounts to share applications with as many as 20 devices without providing government-issued identification or paying a registration fee.
Google also provides an advanced installation process for power users who want to install applications from unverified developers.
Its published rollout calls for verification requirements to begin September 30, 2026, on certified Android devices in Brazil, Indonesia, Singapore and Thailand, with broader expansion planned beginning in 2027.
Those changes reduce some of the concerns raised by the original proposal.
They do not eliminate the broader question:
What happens when a government gains access to a technical chokepoint capable of identifying speakers or limiting distribution?
Apple's own transparency report raises another question
Apple publishes information about government demands involving applications.
Its 2025 App Store Transparency Report says 2,045 applications were removed worldwide subject to government takedown demands.
Russia accounted for 1,213.
Mainland China accounted for 196.
The United States accounted for 39.
But Apple says all 39 U.S. applications in that category were ByteDance applications affected by a federal law and later restored in February 2025.
ICEBlock is not among those 39.
That does not establish that Apple concealed the ICEBlock removal.
Apple said at the time that it acted after receiving information from law enforcement and concluded that the application violated App Store rules.
That may place the removal in a different reporting category.
But it leaves an important public-record question:
Where, exactly, is a government-prompted removal like ICEBlock recorded in Apple's transparency system?
The answer matters because transparency reports are supposed to help the public understand when governments influence what private platforms allow people to access.
The authoritarian warning sign
The warning sign is not simply that a particular application disappeared.
Nor is it whether a reader likes or dislikes ICEBlock.
The larger issue is whether government power can create a repeatable pathway for suppressing lawful expression without openly banning it.
Consider two situations.
In the first, Congress passes a law declaring that Americans cannot publish certain information about government activity.
The constitutional issue would be visible immediately.
In the second, government officials contact a private company controlling distribution and tell it that material needs to disappear.
The result can look like a private corporate decision.
That appearance does not resolve the constitutional question.
We still have to ask:
- Who initiated the contact?
- What was requested?
- What authority did the official possess?
- Was a threat made or implied?
- Could the company reasonably fear prosecution, investigation or regulatory consequences?
- Did the company make an independent decision?
- Or did government pressure determine the outcome?
Those questions separate legitimate government advocacy from unconstitutional coercion.
The comparison with authoritarian governments requires care
Governments including China and Russia have long used restrictions on applications, developers, internet services and online platforms to control access to information.
That does not mean one American application removal makes the United States equivalent to China or Russia.
It does not.
The countries have different laws, institutions, courts and degrees of government control.
The useful question is narrower:
Is the United States beginning to use mechanisms that can also be used by governments to restrict access to information?
If it is, the next questions are about scale, authority and safeguards.
How frequently is it happening?
Who approves it?
Can companies refuse?
Can courts review it?
Can the public find out?
Those are democratic-accountability questions, not partisan ones.
Intent and effect are different questions
Government pressure can weaken a safeguard even when the motive behind the pressure has not been established.
That distinction is essential.
Evidence showing that officials pressured a company establishes something about what happened.
It does not automatically prove why they did it.
Evidence of a broader political censorship strategy would require more.
Internal directives could establish it.
Communications identifying political critics for suppression could establish it.
A repeated pattern of targeting disfavored viewpoints could help establish it.
Statements tying government punishment or benefits to compliance could help establish it.
Without that evidence, motive should not be presented as fact.
The effect and the intent are separate questions.
What the public still does not know
The available record establishes government intervention in the removal of ICE-tracking applications and provides evidence strong enough for a federal judge to issue a preliminary injunction in a related case.
It does not establish the full scale of the practice.
The public still does not know how many comparable requests federal agencies have made to Apple, Google, Facebook or other technology companies.
We do not know how many were refused.
We do not know how many were accepted.
We do not know how often prosecution, investigation or regulation was discussed.
And we do not yet know whether these episodes represent isolated disputes or pieces of a broader government practice.
Those are questions the public record should be able to answer.
The missing safeguard
Constitutional protections do not enforce themselves.
Someone has to challenge the government.
A developer may have to sue.
A company may have to resist.
A journalist may have to find the records.
A judge may have to intervene.
Congress may have to investigate.
The public must be able to learn what happened.
When those safeguards function, government power can be challenged.
When they weaken, government pressure becomes harder to see and easier to repeat.
That does not prove an authoritarian plan. It does tell us where to look for one.
The larger question
The First Amendment was written when there was no App Store.
There was no Google Play.
There were no social-media platforms capable of deciding almost instantly whether information could reach enormous audiences.
The technology changed.
The constitutional problem did not.
Government officials cannot escape First Amendment limits simply by using private intermediaries to accomplish the suppression.
The Supreme Court has said so.
And in 2026, a federal judge applied that principle to government pressure involving ICE-tracking speech.
The larger issue therefore goes far beyond one application.
A communications system concentrated in powerful private companies creates chokepoints.
Those chokepoints can protect people.
They can stop malware.
They can enforce legitimate safety rules.
They can remove unlawful material.
And they can become places where government pressure reaches speech without the government ever passing a censorship law.
That is why the public should watch the mechanism, not merely the individual controversy.
Is America better yet?
The answer cannot depend on whether we agree with the speech that disappears.
A constitutional protection means very little if it protects only speech government officials approve.
The real test comes when speech is inconvenient.
When it angers officials.
When it challenges government policy.
When powerful people want it gone.
At that moment, the important question is not simply:
Who removed it?
It is:
Who made them do it?
And if government power supplied the pressure, one more question must follow:
By what lawful authority?
Sources
Supreme Court of the United States
National Rifle Association of America v. Vullo, 602 U.S. 175 (2024), opinion issued May 30, 2024.
U.S. District Court for the Northern District of Illinois
Rosado et al. v. Bondi et al., No. 1:26-cv-01532, Memorandum Opinion and Order granting preliminary injunction, April 17, 2026.
Electronic Frontier Foundation
Aaron v. Bondi, case page and court filings, reviewed August 18, 2026.
Electronic Frontier Foundation
“Application Gatekeeping: An Ever-Expanding Pathway to Internet Censorship,” November 3, 2025.
Apple
2025 App Store Transparency Report, published 2026.
Google, Android Developers
Android Developer Verification documentation and rollout guidance, reviewed August 18, 2026.
U.S. Department of Justice
Attorneys General of the United States, Pamela J. Bondi, 2025-2026.
I write for you,
— Roberto
isamericabetteryet.org