Google Pushes AI Regulation with Rules Favoring Industry

JAKARTA — Google is once again pushing for AI regulation, but with a design that still leaves plenty of room for industry to move. The company, as reported by The Verge, is asking the United States government to take a “middle path” that it says can protect innovation without letting artificial intelligence run wild.

The request is straightforward. Google wants rules, but not rules that slow down a fast-growing business that is already widely used and fiercely contested by major players.

Google’s version of AI regulation: rules, but room for business

In a blog post by Google president Kent Walker and a 21-page policy paper titled A Pragmatic Approach to AI Governance in America, Google says the debate over AI governance has become trapped in a false choice: too strict or no rules at all.

Walker wrote that there is a “middle path” built on a pragmatic, evidence-based approach. Google even proposes a federal frontier AI watchdog called FARO. The body, according to the company, could oversee the most advanced AI models without shutting the door on AI use that has already spread into daily life.

It sounds tidy on paper. But the real question sits right there: who writes the guardrails, who holds the power, and who benefits most when those guardrails take effect.

Google links the FARO model to other bodies that are described as independent but still operate under government oversight, such as the North American Electric Reliability Corporation, the Financial Industry Regulatory Authority, and the American Medical Association. The pattern is clear. Industry helps shape the standards, while the state watches from above.

For Google, that structure offers balance. For critics, it can look like industry regulating itself under a public oversight label.

What Google is really asking for in AI regulation

In the same paper, Google also pushes practical rules for AI platforms. AI models, it says, should be required to give consistent warnings, filter sexual or romantic content, avoid claiming to be human, and avoid encouraging emotional dependence.

The list sounds reasonable. But the internet has taught a different lesson. Rules that look strict on paper often become formality, while misinformation, emotional manipulation, and platform abuse slip through other gaps.

Google also addresses the use of public web data to train models. The company argues that using such data is a transformational, non-expressive use, similar to an art student being inspired while walking through a gallery. In many countries, Google says, that approach should stay protected through fair use or text-and-data-mining exceptions.

At this point, the AI regulation debate is no longer about whether AI may be used at all. What is being fought over is the limit. How far can a company collect data, train models, and then sell services built on an open ecosystem?

That is where Google’s interests meet the interests of other businesses. For a company that controls infrastructure, data, and product distribution, “balanced” rules can become a comfortable fence. Not a fence meant to restrict them too much. A fence meant to keep the market growing.

Why this AI regulation debate matters to the public

The problem is that AI is no longer a lab issue. It has entered web search, document writing, customer service, education, images, voice, and office work that once felt safe. As AI models spread, the impact reaches ordinary users: what they read, what they trust, and which jobs slowly change shape.

That is why AI regulation is not just a matter for tech companies and officials in Washington. The rules chosen will shape safety standards, data protection, copyright, and the future of creative workers whose income is being squeezed by automation.

Google itself acknowledges the need for retraining programs. In the paper, the company encourages incentives for firms that run retraining efforts. It sounds like a neat compromise: AI keeps moving, and part of the social cost is covered through retraining.

But the larger question remains. Can retraining move fast enough to catch up with the changes AI brings? And who pays for the transition when the market moves faster than worker protection systems?

In the United States, that answer is still being debated. In many other countries, including Indonesia, the basic questions are not even finished yet: who oversees the models, what data can be used, and how user rights are protected as AI is used more widely in public services and everyday business.

Google is not the only company talking about rules. OpenAI and Anthropic have also called for binding regulation, although that tone changes once policy starts touching their own businesses. The Verge notes a repeated pattern: the industry pushes for oversight, then pulls back when that oversight gets too close to commercial interests.

That is why Google’s “middle path” proposal should be read carefully. Not rejected outright, but not accepted at face value either. Behind the word pragmatic, there is always a market calculation.

Google wrote, “We can have rules that allow us to keep doing what we are doing.” The sentence sounds simple. And that is exactly where its message is clearest.

Quick summary:

1. Google supports AI regulation, but wants an oversight model that remains friendly to business.

2. The company proposes FARO, a federal body to supervise frontier AI in a structure similar to other industry-linked watchdogs.

3. The debate matters because it could shape safety, data, copyright, and the future of work in many countries, including Indonesia.

Short FAQ:

What is Google’s main proposal? AI rules should remain in place, but be designed so innovation and large businesses are not slowed down.

Why is the proposal controversial? Because a “middle path” can provide public oversight while still favoring the industry that helps write the rules.

Who reported this? The Verge, citing Kent Walker’s blog post and Google’s policy paper.