Trump can rename AI. Europe’s law still calls it artificial.
Trump’s objection to “artificial intelligence” is that artificial sounds fake. European law uses the word to identify the technology it regulates. The clash is one of terminology: a presidential rebrand does not rewrite the EU’s AI Act or demonstrate a new level of machine capability.
At the UN General Assembly on 22 September 2026, Trump announced that the US would replace artificial intelligence with super intelligence. That proposal raises a practical question for readers in Europe: what changes when the same technology acquires a different political label?
The objection driving the renaming is that artificial makes the technology sound counterfeit. Yet the word also describes how something is made, as in artificial light. The established term artificial superintelligence retains that distinction even at the highest level of capability.
The same distinction runs through the EU’s legal vocabulary. The AI Act, Regulation (EU) 2024/1689, is published in all 24 official languages of the Union. Each language’s term for AI marks the technology as made, built or artificial:
- BGизкуствен интелект
- HRumjetna inteligencija
- CSumělá inteligence
- DAkunstig intelligens
- NLkunstmatige intelligentie
- ENartificial intelligence
- ETtehisintellekt
- FItekoäly
- FRintelligence artificielle
- DEkünstliche Intelligenz
- ELτεχνητή νοημοσύνη
- HUmesterséges intelligencia
- GAintleacht shaorga
- ITintelligenza artificiale
- LVmākslīgais intelekts
- LTdirbtinis intelektas
- MTintelliġenza artifiċjali
- PLsztuczna inteligencja
- PTinteligência artificial
- ROinteligență artificială
- SKumelá inteligencia
- SLumetna inteligenca
- ESinteligencia artificial
- SVartificiell intelligens
Finnish makes the construction especially clear: tekoäly combines teko, made, with äly, intellect. Estonian uses tehis. Across these different forms, the terminology preserves a distinction between constructed and naturally occurring intelligence.
A new label does not rewrite a definition
That vocabulary also has a practical role. Article 3 of the AI Act defines AI system, and the regulation’s obligations apply according to its definitions and scope. A change in US government terminology does not amend those definitions. The US NIST AI Risk Management Framework, a voluntary framework, also uses artificial intelligence.
For organisations working across jurisdictions, the implication is to make the terminology explicit in policies, procurement documents and contracts. Calling a system super intelligence does not determine its capabilities or whether it falls within the AI Act.
Consider a procurement brief that asks for a “superintelligence solution”. The phrase does not tell a reader which tasks the system performs, how its performance was evaluated, or what human oversight it needs. Specifying those properties would make the brief more useful, whichever headline term the supplier prefers.
The terminology clash is therefore a communication problem, not evidence that using the word superintelligence is itself unlawful. A product label, a research claim and a legal definition answer different questions. Keeping those questions separate helps readers assess each one.
The name should follow the capability
A label can express a political preference, an investment ambition or a legal category. Those uses should not be mistaken for evidence of technical performance.
Artificial remains useful because it describes the technology’s origin. General and super make claims about its capabilities. Clear communication should preserve those distinctions: describe demonstrated abilities, state the limits, and reserve claims of superintelligence for evidence that meets the definition.