Political Tech
Microsoft's president describes a regulatory environment built on informal pressure and unspoken expectations — and that should alarm anyone who cares about accountability.
NewsOnScale Staff
July 31, 2026
When one of the most powerful technology executives in the world tells a major financial publication that his industry is being regulated without transparent or complete rules, the temptation is to read it as lobbying. Dismiss it as a billionaire asking referees to put the rulebook away. But Brad Smith's remarks to Fortune this week deserve a harder look, because the problem he is describing — if accurately characterized — is not good for Microsoft, and it is not good for anyone else either.
Smith's framing centers on a regulatory environment where expectations are communicated informally, enforcement priorities shift without public notice, and companies are left to interpret federal intent through a combination of agency guidance, congressional signals, and quiet conversations in Washington corridors. Whether or not you trust Microsoft's motives in surfacing this critique, the underlying condition he describes is real and documented.
## Governance by Implication
Since the Biden administration's 2023 Executive Order on AI and the subsequent unraveling of several of its implementation timelines under the current administration, the federal landscape for AI governance has been defined less by statute than by ambiguity. The major sectoral regulators — the FTC, the CFPB, the EEOC, the FDA — have each staked out partial jurisdiction over AI applications in their domains. But no unified framework exists. No single agency has clear primary authority. And Congress has not passed comprehensive AI legislation.
What fills that vacuum is informal power. Agencies signal priorities through enforcement actions and warning letters. White House offices issue voluntary commitments that carry implicit pressure. Companies that engage cooperatively receive different treatment than those that don't — but none of those distinctions are written into binding law that outside parties can review, challenge, or invoke.
This is the environment Smith is describing. And for all the self-interest embedded in a large corporation calling for clearer rules, the critique lands.
## Who Gets Hurt by Opacity
The companies best equipped to navigate informal regulatory environments are the largest ones — precisely because they have the legal teams, the lobbyists, and the executive relationships to decode unwritten expectations. Smaller AI developers, startups, and non-U.S. competitors operating in American markets do not have those advantages. Informal governance systematically favors incumbents.
Civil society fares even worse. Advocacy organizations, researchers, and affected communities cannot FOIA an informal conversation. They cannot challenge a policy that was never written. They cannot hold an agency accountable for enforcing a standard that was never publicly articulated. Opacity in governance is not neutral — it concentrates power in the hands of those who already have access.
This matters especially on NewsOnScale's core beat. AI agents are now being deployed in hiring pipelines, benefits determinations, content moderation, and political advertising. The question of who is responsible when those systems cause harm — and what rules they were supposed to be following — cannot be answered by pointing to a handshake agreement between an agency official and a corporate compliance team.
## The Brookings Parallel
It is worth noting that this week also brought a renewed call from the Brookings Institution for Congress to pass a federal AI governance law — a separate story, but a related signal. The chorus of voices arguing that informal, fragmented oversight is insufficient is growing. It now includes traditional civil liberties advocates, industry executives with obvious conflicting interests, and mainstream policy institutions. When that coalition converges on the same diagnosis, even if for different reasons, it is usually because the diagnosis is correct.
## What Accountability Requires
Transparent governance means published rules, open comment periods, documented enforcement rationales, and mechanisms for public challenge. It does not mean deregulation. Smith's complaint about incomplete rules should not be read as a request for fewer rules — it should be read as an argument for rules that are actually written down.
If Washington is serious about governing AI in the public interest, the first step is making the rules legible to the public. That is not a concession to industry. That is the minimum condition for democratic accountability.