The GUARD Act Isn't Targeting Dangerous AI – It's Blocking Everyday Internet Use

Summary

Congress is advancing the GUARD Act, legislation intended to protect minors from harmful AI systems but written with sweeping definitions that extend far beyond high-risk applications. The bill classifies any system generating non-pre-written responses as an 'AI chatbot' and labels tools designed for conversational interaction as 'AI companions,' triggering mandatory age verification and access restrictions. This would impact not just specialized AI applications but everyday tools like search engines, homework helpers, and customer service chatbots.

The practical effect would force companies to choose between implementing invasive age verification systems using government IDs or third-party checkers, or blocking all minors from their services entirely. Facing regulatory uncertainty and steep penalties, most companies would likely opt to block younger users rather than navigate vague compliance boundaries. A high school student seeking homework help, a teenager resolving a customer service issue, or any young person using an AI-enhanced service would encounter age barriers.

While the concerns motivating the GUARD Act are legitimate—troubling reports of harmful AI interactions with vulnerable users deserve attention—the proposed solution is disproportionately broad. The article argues that targeted safeguards and enforcement against bad actors would better address specific harms than industry-wide restrictions that affect all users and extract privacy costs from everyone.

• Companies facing regulatory ambiguity will likely choose blanket access restrictions for minors rather than navigate vague compliance boundaries
• Targeted safeguards and enforcement against specific bad actors may be more proportionate than broad restrictions on conversational AI