The GUARD Act Wants Your Government ID for Chatbots

To protect children from AI, Congress voted unanimously to require identity verification for every adult who uses it. The EFF calls it a backdoor to universal online ID checks.

A fingerprint scan displayed on a digital screen with blue lighting

TL;DR:

  • What: The GUARD Act (S. 3062) requires identity verification, government IDs, financial records, or biometric scans, before anyone can access AI chatbots
  • Who voted for it: The Senate Judiciary Committee advanced it 22-0 on April 30, 2026. Bipartisan sponsors: Sen. Josh Hawley (R-MO) and Sen. Richard Blumenthal (D-CT)
  • What it covers: Any system generating "adaptive, human-like responses", chatbots, tutoring apps, customer service bots, AI search tools
  • The real problem: Creates a massive database linking every American's real identity to their AI usage. Constant reverification required. Anonymous use becomes impossible
  • Who gets blocked: Over half of US teens use chatbots for homework. All of them would lose access. Adults without government ID or bank accounts get locked out too

What the GUARD Act Actually Requires

Senator Josh Hawley introduced the Guidelines for User Age-verification and Responsible Dialogue Act in early 2026. Senator Richard Blumenthal co-sponsored. The Senate Judiciary Committee passed it unanimously on April 30, 2026, not a single dissenting vote. [1]

The bill bans anyone under 18 from accessing "AI companions." To enforce that ban, it requires every user, adults included, to prove their identity before they can use any qualifying AI system. The accepted methods: government-issued photo ID, financial records, or biometric identifiers like facial scans. [2]

This isn't a one-time check. The bill mandates "ongoing verification," meaning companies must continuously confirm you are who you say you are. Every 30 minutes, the chatbot must remind you it's not human. It cannot represent itself as a licensed professional. It cannot discuss anything that "risks encouraging minors to engage in sexually explicit conduct." [3]

Penalties: $250,000 per violation, enforceable by both federal and state officials. [4]

The Scope Is Staggering

The original version of the GUARD Act captured "nearly every AI-powered chatbot or search tool," according to the EFF. [5] Congress narrowed it. The new version targets "AI companions", systems designed to simulate emotional or interpersonal interactions, or that present a "persistent identity, persona or character." [4]

Sounds focused. It's not.

That definition sweeps in tutoring apps that encourage students. Language learning tools with persistent AI personas. Customer service bots designed to empathize with frustration. Mental health chatbots. Writing assistants with personality. The EFF warns that even "modern customer service systems designed to empathize" could fall within the bill's definition. [4]

A February 2026 Pew survey found over half of US teens use chatbots for homework help. Every single one would be locked out. [3]

Fight for the Future went further: the bill's verification requirements could apply to "everyone from internet service providers to anyone who runs a blog with a comment section." [3]

The Surveillance Infrastructure Nobody's Talking About

Here's what the "child safety" framing obscures: the GUARD Act would create the largest identity-linked behavioral database in American history.

Think about what it means to require government ID verification for AI chatbot access. Every question you ask an AI, about your health, your finances, your legal problems, your relationships, your politics, gets linked to your verified real-world identity. Forever.

The EFF's analysis is direct: this creates "databases of sensitive identity information" that become "targets for breaches." Anonymous or pseudonymous use of online tools "becomes harder or impossible." [5]

We already know what happens to these databases. In 2024 alone, identity verification company AU10TIX exposed the identity documents of Uber, TikTok, and X users through an unsecured admin credential. ID.me's facial recognition system, used for government benefits, was found to have a 20% failure rate for darker-skinned users. The more identity data you centralize, the bigger the target.

And once this infrastructure exists, expanding its use is trivial. Today it's AI chatbots. Tomorrow it's social media. Then search engines. Then anything with a text input field.

Who Gets Locked Out

Millions of Americans don't have current government ID. They don't have major bank accounts. They don't have smartphones capable of biometric verification. The EFF notes that the bill's verification requirements create "real risks for privacy, anonymity, and data security" and would exclude: [4]

  • Undocumented immigrants (who use AI translation and legal information tools)
  • Unbanked Americans (7.1 million households, per the FDIC)
  • Domestic violence survivors using pseudonymous accounts for safety
  • Military families stationed overseas without current US documentation
  • Teenagers whose parents specifically want them using AI educational tools
  • Anyone who simply doesn't want their government tracking what questions they ask a computer

The First Amendment Problem

The bill compels speech, every chatbot must deliver government-mandated disclaimers. It restricts speech, chatbots cannot discuss entire categories of content, regardless of the user's age. And it conditions access to information on surrendering your identity to the government. [3]

Courts have consistently held that anonymous speech is protected by the First Amendment. The Supreme Court ruled in McIntyre v. Ohio Elections Commission (1995) that "an author's decision to remain anonymous is an aspect of the freedom of speech protected by the First Amendment." Requiring identity verification to access AI tools, which are increasingly how people access information, directly conflicts with this principle.

The "reckless disregard" standard in the bill's content restrictions is particularly dangerous. A chatbot discussing safe sex practices for a teenager who identifies themselves as 18? That's potentially $250,000 in fines. The incentive is obvious: block everything remotely sensitive. Over-restrict. Make the AI useless rather than risk liability.

Who's Opposing It, And Who Isn't

NetChoice, the tech industry trade group representing Google, Meta, Amazon, and others, formally opposed the bill. Their letter argues it would "stifle innovation" and impose "impractical verification burdens." [6]

The EFF has published two detailed analyses calling it both an "everyday internet use" blocker and noting that even after narrowing, "serious problems remain." [4][5]

Fight for the Future opposes. The ACLU has raised concerns about age verification mandates broadly.

But here's what's alarming: nobody in the Senate opposed it. Twenty-two senators. Zero no votes. In a chamber that can't agree on whether to fund the government, both parties found common ground on requiring your papers to talk to a chatbot. "Think of the children" remains the most powerful override for civil liberties in American politics.

What Happens Next

The GUARD Act now awaits a full Senate floor vote. Given the unanimous committee support, passage seems likely. The House would need to pass a companion bill.

If enacted, the practical effects would be immediate:

  • Every major AI provider would need to implement identity verification within the compliance window
  • Smaller developers would face an impossible choice: build expensive verification systems or shut down
  • Open-source AI tools hosted in the US would need to comply or relocate
  • The identity verification industry, already a multi-billion dollar market, would explode

The $250,000-per-violation penalty ensures overcompliance. Companies won't risk it. They'll verify everyone, block edge cases, and store your identity data in databases that will inevitably be breached.

What You Can Do

  • Contact your senators. The bill passed committee but hasn't hit the floor yet. This is the window
  • Support the EFF's campaign. They're leading the opposition and tracking the bill's progress [4][5]
  • Use the language that matters. This isn't a "child safety bill." It's an identity verification mandate for all Americans. Frame it that way
  • Watch for companion bills. State-level versions are likely coming. California, New York, and Illinois already have age verification proposals in various stages

Sources

  1. Global Policy Watch, Senate Judiciary Committee Advances GUARD Act (May 2026)
  2. IBTimes, New Bill Would Require Americans to Submit ID or Face Scan to Use AI Chatbots
  3. Reason, How a Bill Banning AI Companions for Kids Could Usher in Widespread ID Checks Online (May 4, 2026)
  4. EFF, Congress Narrowed the GUARD Act, But Serious Problems Remain (May 2026)
  5. EFF, The GUARD Act Isn't Targeting Dangerous AI, It's Blocking Everyday Internet Use (April 2026)
  6. NetChoice, Letter of Opposition to the GUARD Act