United States

No federal, state, or local authority ordered a political internet shutdown or broad throttling event between June 2025 and May 2026. Independently collected network measurements did not identify a nationwide connectivity restriction. [10][37]

The record matches the top anchor. Provider outages are assessed under reliability and resilience, not counted as government shutdowns.

OONI data do not show systematic state filtering of news or political websites. The federal TikTok divestiture law and state age-verification laws create targeted routes to disabling a service or conditioning access; enforcement of the TikTok prohibition was repeatedly delayed during the period. [10][12][13]

Lawful information remained broadly reachable, but the legal power to disable a major platform and identity-related access burdens are more than theoretical limitations.

Government removal requests are usually tied to identified content and legal process, and major platforms publish some request data. Google recorded 228 U.S. government removal requests in the first half of 2025. Disclosure is incomplete across agencies, and official pressure directed at media and platform decisions was a concern during the period. [15][33]

The practice is not a centralized censorship program, but fragmented transparency and administrative pressure fall short of the strong-review anchor.

Section 230 continues to provide broad protection for hosting third-party content and for voluntary moderation. Federal exceptions, state laws, and age-assurance duties create litigation and compliance incentives that can lead services to remove lawful material or restrict access. [14][13]

The safe-harbor baseline is strong. The growing patchwork of exceptions and duties creates a limited but credible over-removal risk.

VPNs, Tor, encrypted DNS, foreign news services, and independent hosting are lawful and readily available. The research found no government program that blocks these tools for the general public. [10][9]

The legal and technical conditions match the top anchor, with no contrary government restriction documented during the assessment period.

The First Amendment applies fully to online speech and publication. Courts regularly review government action, protect editorial discretion, and require viewpoint-neutral treatment of the press, including in politically sensitive cases. [12][13][19]

The legal baseline and practical availability of judicial enforcement meet the top anchor, notwithstanding disputes addressed in the indicators below.

There is no federal criminal defamation or blasphemy offense. Criminal liability is generally limited to categories such as true threats, fraud, unlawful disclosure, and obscenity. Disputed applications at the edge of the true-threat doctrine and broad online-safety laws create some risk for protected expression. [13][14]

Protected criticism is not routinely criminalized. A half-point deduction reflects live disputes over vague or expansive applications rather than a general pattern of prosecution.

The U.S. Press Freedom Tracker documented at least 32 instances in 2025 in which journalists were detained or charged while doing their jobs; most were released without charges or saw charges quickly dropped. During this assessment period, prosecutions involving Don Lemon and Georgia Fort remained contested, while journalist Mario Guevara was detained in June 2025 and later deported after local charges were not pursued. [16][18][19][20]

These are multiple credible cases, not isolated administrative mistakes. Some arose from alleged conduct rather than publication itself, and the charges remain subject to court review; the score does not presume guilt or improper motive in any pending case.

Press-freedom monitors recorded more than 150 assaults on journalists during protests in 2025, including allegations involving law-enforcement officers. Federal prosecutors also brought cases for threats against reporters and for an assault on a conservative journalist, showing that remedies function in some cases. [17][21]

The volume and recurrence are inconsistent with an isolated-problem score. Prosecutions across viewpoints and the absence of a nationwide policy of violence weigh against a lower rating.

Americans use a large and diverse set of news outlets, platforms, nonprofit organizations, petitions, and campaign tools. Harassment, local-news contraction, concentrated ownership in parts of the media market, and the enforcement cases described above impose unequal costs and can deter participation. [17][14]

The civic space is active and pluralistic, but sustained pressure on reporters and vulnerable speakers is a material limitation rather than a rare exception.

The United States still lacks one comprehensive federal consumer privacy law. Federal rules are sector-specific, while at least 19 states have adopted comprehensive statutes with different coverage, exemptions, enforcement powers, and consumer rights. [22]

The patchwork creates real rights for many residents but leaves protection dependent on location, sector, and enforcement capacity. It does not meet the comprehensive-framework anchor.

Ordinary criminal interception generally requires a warrant, and foreign-intelligence surveillance operates under statutory and court oversight. Section 702 does not require an individualized order for each non-U.S. target; ODNI reported 7,413 FBI queries using U.S.-person identifiers in the covered 2025 reporting period. [23][24][25]

Query reforms, audits, reporting, and the FISA Court are meaningful safeguards. The breadth of collection and the absence of prior individualized review for Section 702 keep the score at 2.5.

There is no general federal real-name rule, SIM registration mandate, or universal data-retention law. Agencies may obtain commercially available information under internal rules; CBP conducts basic device searches at the border without individualized suspicion, and a 2026 visa-vetting policy directs covered applicants to make social-media profiles public. [26][27][28]

Formal anonymity remains widely available, but commercial-data access, border authority, and compelled visibility for visa applicants materially weaken it for affected people.

Federal policy restricts government use of commercial spyware associated with security or human-rights risks, and the Treasury has sanctioned vendors. Reporting during the period showed that ICE had approved a commercial spyware tool for specified investigations; DHS later said the agency had no current relationship with the named vendor. Public records did not establish political targeting, but the warrant and targeting safeguards were not fully disclosed. [30][31]

The evidence supports concern about procurement and transparency, not a finding of systematic political misuse. Formal safeguards and sanctions prevent a lower score; incomplete public accountability prevents a higher one.

Strong end-to-end encryption, anonymous browsing, and security research remain lawful, and there is no general statutory backdoor or key-escrow requirement. The Justice Department continues to seek lawful-access solutions, while border authorities may request device credentials from travelers. [29][27]

The baseline matches the top anchor, with a half-point deduction for border-search exposure and sustained government pressure for exceptional access.

Federal and state courts can enjoin executive action, review platform and speech laws, and publish reasoned decisions. National-security and immigration proceedings provide less public access, and relief can arrive after the practical harm has occurred. [12][13][19]

Judicial review is independent and consequential in ordinary practice. Limited secrecy, delay, and compliance disputes justify a half-point deduction.

The FCC and FTC act under public statutes and remain subject to judicial review. The 2025 removal of protected FTC commissioners without relying on the statutory for-cause standard, together with political disputes over FCC media investigations, placed agency independence under direct pressure. [32][11]

Legal process and review remain available, but leadership security and selective-enforcement concerns are substantial enough to fall below a generally independent score.

Agencies publish FOIA reports, surveillance statistics, court opinions, regulations, and some request data. The federal government received about 1.7 million FOIA requests and processed about 1.6 million in fiscal 2025; significant backlogs, classification, and fragmented reporting remain. [33][34][24]

Core categories are public, but the record is neither consistently timely nor complete enough for the higher anchor.

People and organizations can challenge government action in federal court, use administrative appeals, and sometimes obtain injunctions, suppression, records, or damages. Cost, delay, standing rules, surveillance secrecy, and immigration status can make the remedy inaccessible or too late. [19][34][23]

Remedies change outcomes in practice, but they are not consistently prompt or accessible to everyone affected.

Notice-and-comment rulemaking, congressional hearings, technical consultations, and civil-society participation are normal features of U.S. internet policy. NTIA formally supports multistakeholder governance. Some major decisions during the period moved through accelerated executive or litigation-driven processes with limited opportunity to shape the outcome. [35][12]

Participation is routine and consequential in many proceedings, but it is uneven across national-security, immigration, and executive actions.