Google proposes AI regulation that would allow it to continue current practices with minimal disruption.
For more than three years, AI executives have called for government regulation of their industry—until such oversight threatened their business. OpenAI and Anthropic led this pattern, with Anthropic CEO Dario Amodei publicly calling for "binding regulations" in June 2026, only to push back when his latest models were suspended.
Google, which has also advocated for AI regulation, now seeks to clarify its position and propose a "middle way" designed to favor its interests.
"The debate over AI governance is stuck in a false choice between over-regulation and no regulation," said Google president Kent Walker in a blog post. "There is a middle way: A pragmatic, evidence-based approach that recognizes the unique challenges and opportunities of both frontier AI and widely-deployed AI applications."
Though Walker does not explicitly define "over-regulation," the reference appears to be the recent ban on Anthropic's Fable 5 and Mythos 5 models.
Google's 21-page policy paper, "A Pragmatic Approach to AI Governance in America," proposes a federally overseen frontier AI regulatory organization (FARO), modeled after industry-funded bodies like the North American Electric Reliability Corporation, the Financial Industry Regulatory Authority, and the American Medical Association—each ostensibly independent yet subject to government oversight.
This "middle path" sits uneasily with a decade of existential warnings from tech leaders. If AI truly poses catastrophic risk, one might expect regulation akin to lead or asbestos. Yet Google argues that "AI platforms should be required to take reasonable measures to feature persistent disclaimers, filter out sexually explicit or romantic content, avoid claims the model is a person, and not promote emotional dependency."
Similar safety measures have failed elsewhere. Section 230 of the Communications Decency Act has shielded platforms taking performative steps, yielding free speech benefits alongside unchecked misinformation and social incitement.
The middle path for AI governance already exists: an acceptable level of chatbot suicide promotion, non-consensual synthetic imagery, copyright infringement, model bias, indemnified errors, and guidance toward harms.
Google frames datacenters as a negotiation with communities: "The question is not datacenters or no datacenters, but how to build datacenters the right way, responsibly and in partnership with communities." Yet for many communities, the answer remains categorical opposition. If one position unites the political spectrum now, it is resistance to datacenters.
On copyright, Google's position reads less as compromise and more as permission to continue. The paper argues that "using publicly available web data for training models is a transformative, non-expressive use—like an art student taking inspiration from walking through a gallery—that should remain protected under fair use in the U.S. and text-and-data-mining exceptions abroad."
The analogy falters. Google's operation more closely resembles an art student who controls tourism referrals capturing the Louvre's entire collection, then profiting by selling access and laundered variations that discourage visits to the original institution. Meanwhile, the displaced creators receive only peripheral job-retraining incentives.
Google requests a middle path, yet AI industry lobbying has surged 340 percent since 2023, suggesting the middle road is being paved toward industry-favorable outcomes.