How AI Became 'Super Intelligence': Trump's Polls Day by Day

The short answer: three polls in four days

Artificial intelligence became "Super Intelligence" through three social media polls run between September 19 and September 22, 2026. The final one pitted "Superior Intelligence" against "Super Intelligence" on Truth Social, and Super won 65% to 35%, according to Gizmodo.

The detail most coverage skips: "Super" was never on the original ballot. The first poll offered Superior, Extreme, and Supreme. Super appeared on day three as a late addition — and beat the two-time winner by thirty points. That reversal is the most interesting part of the whole episode, and it comes down to one grammatical fact about the word "superior."

For what was actually announced at the UN and what it means for organizations, see our breakdown of the UN speech where AI was renamed SI. This article covers only how the name was chosen.

Timeline: September 19–22, 2026

All dates are US Eastern Time. Only events confirmed in reporting are listed.

Day three is the hinge. The first two rounds asked voters to choose among options they were given. The third round introduced a new option mid-process — the format stayed a vote, but the ballot itself was redesigned.

The original three names, and what each one actually signals

All three original candidates replaced the adjective "artificial" rather than the noun. From the September 19 post:

Many people think that the words "Artificial Intelligence" are inaccurate, and very ineloquent, relative to AI, or Artificial Intelligence. A far more elegant and accurate description of this new phenomena would be Superior Intelligence (SI) or, Extreme Intelligence (EI) or, Supreme Intelligence (SI)...

Source: IBTimes UK, transcribing the Truth Social post. Truth Social itself sits behind a bot check, so we cite the outlets that reproduced the text rather than the post page.

Notice that three of the four candidates abbreviate to "SI." Only Extreme produced a distinct set of initials. Since the shorthand could not distinguish the options, the poll was effectively asking which adjective people liked the sound of.

Why "Supreme Intelligence" was cut on day one

The reason was stated outright in the post announcing the runoff:

'Supreme Intelligence,' probably because of its relationship to the Supreme Court, is losing badly to both 'Superior' and 'Extreme Intelligence.' Therefore, we are going to take 'Supreme Intelligence' OUT, deleting it as a qualifier, and let you vote for the Final Two: Superior Intelligence, or Extreme Intelligence. A fresh Vote begins now!

Source: Just the News. In American English, "Supreme" is effectively owned by one institution. A name that competes with an existing proper noun loses even when the literal meaning fits — a failure mode any naming team will recognize.

Why "Superior" won twice and still lost 65–35

Superior Intelligence topped the three-way poll and won the runoff. Then, against a newcomer, it lost by thirty percentage points to a word that differs by two letters. The gap is best explained by what the two words do grammatically.

"Superior" is a comparative. "Super" is not.

Superior comes into English as a comparative form (from Latin superus, "above"), and it still behaves like one: you are superior to something. Attach it to machine intelligence and the implied second term is human intelligence. "Superior Intelligence" quietly names people as the inferior party.

Gizmodo made exactly this point, noting that calling AI "superior" more or less requires calling human intelligence "inferior" — an awkward message for an administration trying to calm fears about AI displacing workers.

"Super" takes no object. As a prefix it means "above," but in everyday usage — superman, supermarket, super easy — it functions as friendly hyperbole. It elevates the thing it modifies without demoting anything else. That asymmetry is the most plausible explanation for a thirty-point swing between two near-identical words.

This nuance largely disappears in translation. In Japanese, for example, both "優れた知能" (superior) and "超知能" (super) read as straightforwardly positive, which is why non-English coverage of the rename rarely explains why Super won.

The chosen term also collides with existing usage: Nick Bostrom's 2014 book Superintelligence defines it as "any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest" — a hypothetical future system, not today's chatbots. We unpack that overlap in our guide to what "super intelligence" means and how it differs from AI, and compare the terms side by side in AGI vs. ASI vs. SI.

Three naming lessons worth keeping

In practice, most name failures we see repeat one of these patterns, and this four-day sequence hit all three:

• Check the initials before the shortlist. Three candidates collapsed to "SI." Listing abbreviations alongside full names would have caught it on day one.
• Search the word before you defend it. Supreme lost to an institution that already owns the term. Run the candidate through a search engine before it reaches a vote.
• Watch for comparatives. Adjectives that imply a ranking create a loser. Ask who your name is quietly placing below your product.

How much should you trust these numbers?

Not very much, as polls. These were opt-in polls of one person's social media followers, not representative samples. IBTimes UK explicitly cautions that respondents should not be treated as a stand-in for US voters.

The figures themselves also vary by outlet and timestamp:

• Superior's day-one share: Gizmodo reported 41.5% on X; Forbes JAPAN reported 42% on Truth Social as of 1:00 p.m. US time on September 19. Different platforms, different snapshots.
• Day-one turnout: Unite.AI recorded 42,206 votes with one day of voting left. No final tally was published.
• The September 20 runoff: reported as narrow, but we could not confirm a percentage split from primary material.
• The 65–35 result: reported by Gizmodo for the Truth Social poll.

We could not retrieve the original Truth Social posts directly — the site requires a bot check. Every quotation here is sourced to an outlet that reproduced the text. We are flagging that rather than implying first-hand access.

Will "SI" actually replace "AI"?

As of September 23, 2026, the announcement covers US government documents only. It carries no force over private companies, other governments, or academic usage, and no executive order or other legal instrument has been confirmed. Trump himself hedged at the UN: "Let's see if I have any power. Maybe I do and maybe I don't."

The realistic near-term outcome is coexistence rather than replacement: federal documents using SI while companies, researchers, and the rest of the world keep writing AI. The practical risk for organizations is not renaming — it is inconsistency when quoting US federal material in contracts, policies, and investor communications. For related US policy context, see our overview of state-level AI regulation in the US.

At Mihata we track AI news back to primary sources before we publish — which is how this article ended up flagging what we could not verify. If your team is working out how to write about fast-moving AI stories without overstating what is confirmed, we would be glad to talk.