Trump Signs Executive Order Replacing 'Artificial Intelligence' in Federal Use
President Donald Trump signed an executive order on September 29, 2026, directing the federal government to stop using the term "artificial intelligence" in official communications. Under the order, executive branch policy websites, policy documents, and press releases will instead refer to "Super Intelligence." The directive amounts to a government-wide vocabulary mandate, an instrument presidents have used before to shape how agencies describe the technologies and risks they regulate.
Trump framed the change in characteristic terms. According to The Verge, the president said, "The word super is the best word of all, and it's the simplest." The remark encapsulates the logic of the order: not a technical reclassification but a rhetorical one, delivered through the most formal mechanism available to the executive branch.
The order does not, based on the reported summary, create new regulatory authority over AI systems or reorganize any federal office. Its reach is linguistic. That narrowness is precisely what makes it worth examining. Terminology sits upstream of budgeting, procurement, research classification, and treaty language. When the world's largest federal research funder changes the words on its forms, the effects ripple outward — to universities, contractors, standards bodies, and allied governments that align their own vocabularies with Washington's.
The shift also arrives at a moment of unusual sensitivity. AI governance has become a matter of national security framing, export controls, and international competition. A naming convention that sounds promotional rather than descriptive may complicate the work of career officials who must explain, in writing, what a system does and what it cannot do.
Which Federal Documents and Channels Are Affected
The order's stated scope covers three categories: official policy websites, policy documents, and press releases. That is a broader footprint than it first appears. Federal policy websites include agency portals that host guidance, rulemaking dockets, and compliance resources used daily by private-sector lawyers and state regulators. Policy documents span strategy papers, memos, and reports. Press releases carry the government's public framing to journalists and markets.
Read next Laika's Wildwood: Stop-Motion Fantasy at TIFF 2026What remains unclear from the summary is how the order treats legacy materials. Agencies maintain decades of archived documents, datasets, and grant records that use "artificial intelligence." Whether those archives are exempted, grandfathered, or subject to retroactive editing will determine how much administrative labor the directive actually requires. Federal records law generally discourages altering archived documents, which suggests agencies will apply the new language prospectively.
There is also the question of statutory language. Congress has passed legislation that uses "artificial intelligence" in defined terms, and the order cannot rewrite a statute by fiat. Agencies implementing those laws will face a recurring drafting problem: how to translate a term of art into a term of preference without creating legal ambiguity. Similar tensions have arisen in past terminology directives, and they are typically resolved through internal style guidance rather than public rulemaking.
Enforcement mechanisms are not described in the reported summary. Historically, terminology mandates of this kind are implemented through the Office of Management and Budget's review of agency communications and through departmental style guides, not through penalties.
Reactions from the Tech and Policy Community
Response has split along predictable lines, with the sharpest criticism coming from researchers who depend on stable terminology. AI governance specialists argue that "artificial intelligence" functions less as branding than as a category boundary. Funding agencies use it to route proposals to the correct review panels. Standards bodies such as the National Institute of Standards and Technology use it to organize risk frameworks. Journals index by it. Change the label at the federal level, and you introduce friction at every handoff point between government and the research community.
Linguists make a related point. Terms of art gain authority through consistent use, and official vocabularies are among the strongest signals available to institutions. When a government adopts a promotional register for a technology it also regulates, the gap between the descriptive and the aspirational can widen. Scholars who study political language have documented this pattern across administrations of both parties: the label chosen in a press release tends to migrate into grant abstracts, corporate filings, and eventually public understanding.
Industry reaction appears more muted. Companies have strong incentives to mirror federal vocabulary when bidding on contracts or seeking regulatory clarity, but they also have brand investments in existing terminology. A firm that sells "AI compliance software" cannot easily rename its product category overnight. Expect a period of dual usage in which vendors describe their offerings one way to customers and another way to federal procurement officers.
Historical Context: How Government Language Shapes Technology Policy
Presidential vocabulary has reshaped technology and science policy before. During the Obama administration, federal agencies issued guidance encouraging the use of "climate change" over "global warming" in official communications, on the reasoning that the broader phrase better captured the range of impacts — shifts in precipitation, sea level, and storm intensity, not just temperature. That guidance was not binding law, yet it propagated quickly through agency reports, grant language, and eventually corporate sustainability disclosures. The terminology shift outlasted the administration that initiated it.
A parallel case sits in cybersecurity. The federal government's long-running effort to standardize terms like "cyber incident" and "breach" through frameworks developed at NIST produced measurable effects on how companies structure disclosure obligations. Vocabulary, in these instances, functioned as quiet infrastructure.
The current order extends a pattern of aggressive executive action on technology. Across 2025 and 2026, the administration issued a series of directives addressing AI development, procurement, and national competitiveness, consistently favoring deregulatory framing and speed over precautionary review. The September 2026 order is continuous with that posture: it removes friction from the government's own language rather than adding it through new rules.
Implications for AI Research, Industry, and International Standing
The practical stakes concentrate in three areas. First, research categorization. Federal grant systems rely on keyword taxonomies to assign proposals to reviewers. If the government's own vocabulary diverges from the vocabulary used across the field's literature, administrative reconciliation costs rise. Academic institutions, which receive a substantial share of their computing and research funding from federal sources, will likely maintain internal terminology while producing federally compliant external documents — a duplication that consumes staff time.
Second, industry compliance and marketing. Companies that sell to government must respond to solicitations written in the government's preferred terms. Procurement language shapes product roadmaps more than most observers assume, because vendors build to the specification documents they must answer.
Third, international alignment. Multilateral discussions on AI safety, including work conducted through the OECD and bilateral science agreements, depend on shared definitions. A unilateral change in the terminology of the largest research funder introduces translation problems into negotiations where precision already is scarce. Allied governments may or may not follow Washington's lead; either outcome creates friction, since divergence in vocabulary complicates joint research programs and mutual recognition of safety testing.
What Comes Next: Enforcement and Practical Impact
The order takes effect immediately according to the reported summary, but the operational timeline will stretch much longer. Agencies must revise style guides, retrain communications staff, and update templates. Federal websites typically number in the hundreds across the executive branch, and each carries archived pages that reference the old term.
Watch three indicators over the coming months. First, whether the Office of Management and Budget issues implementing guidance that defines the boundaries of the new term, including whether "Super Intelligence" covers narrow tools such as document classifiers or only frontier systems. Second, whether Congress reacts, given that existing statutes use "artificial intelligence" as a defined term and legislators from both parties have shown interest in maintaining technical precision in law. Third, whether federal research agencies adjust their grant taxonomies, which would convert a rhetorical directive into a measurable change in how science is funded and catalogued.
For now, the order's most durable effect may be documentation. Every federal press release issued after September 2026 will carry the new vocabulary, and researchers studying the period will be able to date the shift precisely. Governments change words more easily than they change systems. But words, as the historical record shows, tend to travel.
Source: The Verge
