Brad Smith Says U.S. AI Policy Is Broken. He's Not Wrong — But He's Not Neutral Either.

NewsOnScale Staff

July 12, 2026

There is a version of Brad Smith's critique of U.S. AI regulation that is entirely correct. The United States has, in the two years since generative AI became a household concept, produced a patchwork of executive orders, agency guidance documents, voluntary commitments, and state-level proposals that add up to something less than a coherent framework. Companies building AI systems — and the governments, hospitals, schools, and newsrooms deploying them — genuinely do not know what the rules are. That is a real problem.

But Smith is not a neutral observer filing a civic complaint. He is the president and vice chair of Microsoft, one of the largest investors in OpenAI and one of the most consequential actors in the global AI infrastructure market. When he calls for regulatory clarity, it is worth asking: clarity for whom, built around what principles, and enforced by which institutions?

## What Smith Actually Said

Speaking publicly this week, Smith characterized the current U.S. approach to AI governance as regulation without transparent or complete rules — a system that creates compliance burdens without delivering the predictability that businesses or the public need. The framing positions Microsoft as a company willing to be regulated, if only regulators would get their act together.

This is a familiar posture in technology policy. It is not inherently dishonest. Large incumbents often do want some regulation — particularly the kind that raises barriers to entry for competitors, codifies existing practices as compliant, or preempts more aggressive state-level action. The question is not whether Smith is lying. It is whether "clarity" as Microsoft defines it would actually serve the public interest, or primarily serve the business interests of companies already operating at scale.

## The Regulatory Vacuum Is Real

Set aside the messenger for a moment. The substantive problem Smith is pointing at deserves serious attention. Federal agencies have issued guidance on AI use in hiring, lending, and healthcare — but that guidance often carries no enforcement teeth. The executive orders on AI safety issued by the Biden administration created processes and timelines without creating durable legal obligations. Congress has held hearings. Frameworks have been drafted. Very little has been enacted into law.

The result is a system where the most powerful AI developers operate largely on self-certification and voluntary commitments. That is not a foundation for accountability. It is a foundation for liability evasion.

Smaller companies, nonprofits, and public-sector institutions face a different version of the same problem. Without clear rules, they cannot confidently deploy AI tools in high-stakes contexts — or they deploy them without the guardrails that clear regulation might require. The vacuum does not protect people. It mostly protects whoever has the legal budget to navigate ambiguity.

## What Meaningful Clarity Would Actually Require

If the U.S. is going to build AI regulation that works, it will not come from letting the largest AI companies define what "clear" means. Genuine regulatory clarity in this space requires mandatory transparency about how high-stakes AI systems make decisions, independent audit rights for regulators and affected parties, liability frameworks that do not simply transfer risk to end users through terms of service, and enforcement mechanisms with actual teeth.

It also requires Congress to act — not just agencies. Executive guidance can be reversed with an administration change. Voluntary commitments have already proven insufficient. The EU's AI Act, whatever its flaws, established that democratic legislatures can pass binding AI law. The U.S. has not yet demonstrated the same.

## The Accountability Gap

NewsOnScale covers the AI agent economy because that is where power is being restructured right now, often faster than public institutions can track. When a figure like Brad Smith steps into the policy arena calling for clearer rules, that call should be taken seriously — and examined seriously. The public interest in AI regulation is not identical to Microsoft's interest in AI regulation, even where they overlap.

What the U.S. needs is not just clarity. It is accountability. Those two things are related, but they are not the same.