Understanding the Revised GUARD Act: A Guide to Its Pitfalls and Privacy Risks

Overview

The GUARD Act (Generative AI Use Restriction and Data Act) was introduced to restrict minors' access to certain AI systems. After widespread criticism, lawmakers narrowed its focus from broadly covering nearly all AI-powered chatbots and search tools to specifically targeting so-called 'AI companions'—conversational systems designed to simulate emotional or interpersonal interactions. While this narrowing addresses some of the broadest concerns, serious problems remain for privacy, online speech, and parental choice. This guide explains the revised bill's key provisions, its potential impacts, and common misconceptions.

Understanding the Revised GUARD Act: A Guide to Its Pitfalls and Privacy Risks — Source: www.eff.org

Prerequisites

To understand this guide, you should be familiar with:

Basic US legislative process (bills, amendments)
How AI chatbots and conversational agents work
Concept of age verification systems (e.g., government ID, credit checks)
Fundamental privacy and data security principles

Step-by-Step Breakdown of the Revised GUARD Act

1. The Narrowed Definition of 'AI Companion'

The revised bill defines an AI companion as a system that 'engages in interactions involving emotional disclosures' from the user or presents a 'persistent identity, persona, or character.' This excludes general-purpose chatbots like search engines but still encompasses many popular AI tools used by teens, such as character-based roleplay bots or therapy simulators.

Example: A conversational AI that asks 'How are you feeling today?' and adapts responses based on mood would likely qualify. A weather bot that simply reports forecasts would not.

2. Mandatory Age Verification Tied to Real-World Identity

The core requirement remains: companies offering AI companions must implement 'reasonable age verification' that confirms users are adults (18+). Permitted methods include:

Uploading government-issued ID
Providing financial records (e.g., credit card, bank account)
Using age-verified accounts from mobile OS or app stores (e.g., Apple's or Google's age verification)

All these methods are tied to the user's real-world identity, raising privacy red flags. Even if a parent wants a teenager to use an AI companion for therapeutic or creative purposes, the system must perform an identity check on the teen—or the parent must verify their own identity and then supervise, which still adds friction.

3. Impact on Parental Choice and Access

The bill creates significant hurdles for families. Consider these scenarios:

Social skill practice: A teen with social anxiety uses an AI to rehearse conversations. Under the bill, the teen must first pass an age check, which may require personal information that compromises anonymity.
Military family: A deployed parent sets up an AI storyteller for their young child. The parent must verify their own identity (fine), but the child's use still requires identity-linked verification as the system 'presents a persistent character.'
Creative roleplay: A group of teens uses a fantasy RPG chatbot. The platform must block all users under 18 unless they undergo identity verification.

Many families will simply choose not to use these tools rather than submit to invasive data collection.

4. Unclear Definitions and Heavy Penalties

The bill's definitions of 'emotional disclosures' and 'persistent identity' are vague. For example: If a user says 'I'm sad' to a general AI assistant, does that count as an emotional disclosure triggering the rules? The bill leaves room for interpretation, but companies face steep penalties for getting it wrong. This encourages over-compliance: blocking all users under 18 or pushing intrusive verification even when unnecessary.

The liability shift creates a chilling effect on innovation. Startups and small developers may exit the AI companion space entirely, leaving only large corporations that can afford compliance teams and identity verification infrastructure.

Common Mistakes

Mistake 1: Assuming the narrowed scope fixes everything

Many critics focused on the original broad reach; the revision addresses that but introduces its own problems. The core age verification requirement remains privacy-invasive and restricts access for beneficial uses.

Mistake 2: Believing age verification is simple or universal

Millions of Americans lack government ID, stable banking relationships, or consistent access to mobile identity systems. Tying verification to real-world identity excludes these users and creates data security risks (breaches of sensitive documents).

Mistake 3: Overlooking the impact on speech and autonomy

The bill treats AI companions as uniformly harmful, ignoring their legitimate roles in education, mental health support, and social development. By requiring adult-level verification to use any such tool, it undermines the ability of teens to explore ideas, learn social skills, or receive therapeutic interactions anonymously.

Summary

The revised GUARD Act, while narrower, still imposes mandatory identity-linked age verification on AI companions, harms parental choice, leaves critical definitions ambiguous, and exposes companies to heavy penalties that discourage innovation. Privacy and speech concerns persist, and the bill fails to balance protection with access. Lawmakers should reconsider the approach to avoid unintended consequences.