Google backs AI regulation, but only on its own terms

Google, which previously declared broad support for regulating artificial intelligence, has refined its position: the company wants government intervention, but strictly on its own terms.

From support to conditions

For several years, Google's executives publicly called for government oversight of AI development. Pichai, OpenAI's Altman, Meta's Zuckerberg: at various times, all said the industry needs regulation. It looked like corporate responsibility.

There is a rational explanation for this behavior, however. Strict rules and compliance requirements favor big players, which have legal departments, established processes and the resources to run audits. For startups and small teams, the same requirements become an unbearable burden. Regulation turns into a competitive barrier that keeps new players from growing.

Now abstract support for regulation no longer satisfies Google. The company has moved to specific proposals, and they primarily fit its own business model and current market position.

What Google proposes

Among the company's key positions:

• Federal law taking precedence over the rules of individual states: a single national standard reduces operational complexity for a corporation operating in all 50 states
• Regulate applications, not foundation models: Google develops the Gemini family of foundation AI models, so direct regulation at the model-developer level is unfavorable to it
• Voluntary standards instead of strict mandatory rules: self-regulation, in which big players write the rules for themselves
• Transparency without disclosing training data: removes the need to report on the sources used to train models

Each of these points reduces the regulatory burden on Google's core businesses, shifting responsibility onto developers of end applications or onto smaller market players.

A familiar tactic

Google is not a pioneer of this strategy. Every large tech company has discovered the advantages of "the right" regulation at the right moment. Microsoft actively supported AI regulation just when Copilot strengthened its position in the enterprise segment. Meta consistently argues for open models because it has already released Llama as open source and no longer loses a competitive advantage by doing so.

"Companies support regulation when, and in the form that, helps them

lock in their lead," an observation that repeats itself with every major technological shift.

Google now has an additional incentive to claim a favorable regulatory position: Gemini trails ChatGPT in reach among the mass audience. Regulation that slows OpenAI and Anthropic's pace of growth would give Google time to catch up. Sometimes it is easier to back rules that burden competitors than to build a product users like.

What this means

The debate over AI regulation is a continuation of market competition by other means. When Google says "we are for sensible regulation," a specific set of rules written around its business model stands behind that statement. Whose version of "sensible" wins in Washington in the coming years deserves as much attention as new model releases.

Meta has been designated an extremist organization and is banned in Russia.