Google proposes AI regulation body it would help fund

Key Takeaways

• Google proposes a federally overseen, industry-funded AI regulatory organization (FARO) modeled after bodies like FINRA
• AI lobbying has increased 340% since 2023, raising questions about who shapes the 'middle path'
• Google's paper argues web data training is fair use, a position courts are still deciding

Google wants the U.S. government to regulate AI, just not in ways that would hurt Google. The company published a 21-page policy paper this week proposing what it calls a 'middle path' for AI governance: a federally overseen, industry-funded regulatory organization. If this sounds familiar, it should. The model mirrors bodies like the Financial Industry Regulatory Authority, where the regulated industry pays for its own oversight.

Google president Kent Walker framed the proposal as a solution to a 'false choice between over-regulation and no regulation.' The timing matters. Anthropic's CEO Dario Amodei called for 'binding regulations' in June 2026, then pushed back when regulators suspended his company's Fable 5 and Mythos 5 models. Google is now positioning itself as the reasonable adult in the room.

What would Google's proposed FARO actually do?

The proposed Frontier AI Regulatory Organization would sit under federal oversight while drawing funding from AI companies themselves. Google points to precedents: the North American Electric Reliability Corporation for power grids, FINRA for brokerages, the American Medical Association for physicians. Each answers to a government agency but operates with industry money and participation.

Under Google's proposal, AI platforms would face requirements to include persistent disclaimers, filter sexually explicit content, avoid claiming models are human, and discourage emotional dependency. These are, notably, requirements Google already claims to meet. The proposal asks for rules that validate existing practice rather than demanding new constraints.

The copyright question Google wants settled

The policy paper takes a firm position on training data: scraping the public web for AI training should count as fair use. Google compares it to 'an art student taking inspiration from walking through a gallery.' The analogy breaks down quickly. This particular art student controls global search traffic, captured the Louvre's entire collection, and now sells access to derivative images while the original artists lose commissions.

Courts are still weighing multiple lawsuits over AI training and copyright. Google's paper attempts to set the narrative before judges rule. Whether fair use doctrine stretches to cover commercial AI training at this scale remains genuinely unsettled law.

Why the 'middle path' argument rings hollow

For a decade, AI leaders warned that their technology posed existential risks. If those warnings were serious, the regulatory response should resemble how we handle lead or asbestos. Instead, we get proposals for 'reasonable measures' and 'persistent disclaimers.' The gap between the rhetoric and the requested rules is wide enough to drive a datacenter through.

Section 230 of the Communications Decency Act offers a preview. That law shielded platforms from liability if they made good-faith moderation efforts. The result: performative safety theater alongside rampant misinformation. Google's proposed AI framework could produce the same outcome. Some chatbot suicide promotion, some non-consensual deepfakes, some model bias. Hey, they tried.

That 340% increase in AI lobbying since 2023 tells the real story. The 'middle path' is being paved with industry money. When companies call for regulation while spending record amounts to shape it, the request is for favorable rules, not strict ones.

Datacenters: the local fight Google can't lobby away

Google's paper includes a telling admission: 'The question is not datacenters or no datacenters, but how to build datacenters the right way, responsibly and in partnership with communities.' But for many communities, the question is already answered. Opposition to datacenters has become one of the few issues uniting the political spectrum. Water usage, power demands, noise, and local grid strain have turned NIMBY sentiment into organized resistance.

This is where the federal framework Google wants might actually backfire. A FARO could preempt local regulation, letting AI companies bypass community objections. That would accelerate deployment but deepen the political backlash.

Frequently Asked Questions

What is Google's FARO proposal for AI regulation?

Google proposes a Frontier AI Regulatory Organization, a federally overseen but industry-funded body modeled after FINRA and similar organizations. It would set standards for frontier AI and widely deployed applications while allowing companies to participate in rule-making.

Does Google's AI policy paper address copyright concerns?

Yes. Google argues that training AI on publicly available web data should be protected as fair use, comparing it to an art student drawing inspiration from a gallery. Courts are still deciding this question in multiple pending lawsuits.

Why are AI companies asking for regulation now?

Several factors: public concern over AI harms, pressure from the EU AI Act, and the desire to shape rules before stricter mandates emerge. Critics note that companies typically push back when proposed regulations would actually constrain their business models.

How much has AI lobbying increased?

AI industry lobbying has increased 340% since 2023, according to analysis cited by The Register. This suggests significant industry investment in shaping favorable regulatory outcomes.

Source: www.theregister.com