AI rebrand is the shortest honest description of what Washington did to artificial intelligence in the last week of September 2026. On 29 September President Donald Trump signed an executive order telling federal agencies to stop saying “artificial intelligence” and start saying “Super Intelligence”. On the same day he and six technology leaders signed a voluntary safety pact that he called “morally binding”. On 4 October he added a Super Intelligence Force, chaired by Director of National Intelligence Jay Clayton.
TechCrunch’s Equity podcast asked the obvious question on 4 October: can a new name and a non-binding pact fix the image problem that artificial intelligence has with the public? Its hosts were sceptical. Kirsten Korosec called the meeting “a rebranding effort around AI”, and summed up the message as: “AI is going to kill us, it’s scary, it is going to take jobs, but super intelligence is not.”
We have already covered the rename itself in our report on Trump’s super intelligence announcement at the UN, the pact’s four rules in how tech leaders will self-police under the AI safety accord, and the new task force in our Super Intelligence Force explainer. This article asks a different question. Does an AI rebrand ever work, what does the research on names and labels say, and what would actually change how people feel about the technology?
Table of contents
- What the AI Rebrand Actually Changed
- Why the White House Wants an AI Rebrand Now
- Can a New Name Change How People Feel About a Technology?
- What the Non-Binding Pact Does for AI’s Image
- Where the AI Rebrand Could Backfire
- What Would Actually Fix AI’s Image Problem
- What the AI Rebrand Means for UK Businesses
- What to Watch Next in the AI Rebrand
- AI Rebrand FAQ
- References and Further Reading
What the AI Rebrand Actually Changed
The AI rebrand arrived as three separate moves inside two weeks. Each part of the AI rebrand changes less than its headline suggests.
The AI rebrand order changes the name, not the law
Executive Order 14434, “Inaugurating the Era of Super Intelligence”, says modern systems “far exceed what was envisioned when the term ‘Artificial Intelligence’ first came into use”. It orders that “to the maximum extent permitted by law, the executive branch shall use the terms ‘Super Intelligence’ and ‘SI’ in place of ‘Artificial Intelligence’ and ‘AI'”. The executive branch “will not acknowledge” the older terms where the order applies, according to Computerworld’s reading of the text.
Yet the order defines the new term with the old one. Section 3 says Super Intelligence means the technologies covered by the statutory definition of artificial intelligence in 15 U.S.C. 9401(3). The White House science adviser has 60 days, to 28 November, to propose a new legal definition, and changing the statute itself needs Congress.
The scope of the AI rebrand is narrow. The rename covers official correspondence, websites, reports and policy documents. Previously issued regulations, contracts and grants do not need to change. The order says nothing about vendor proposals or product documentation that use the older terms.
The pact adds controls, not consequences
The pact is titled the “White House Accord on Super Intelligence” and subtitled “Joint Commitment on Frontier Responsibilities”. It carries the signatures of Sundar Pichai of Google, Dario Amodei of Anthropic, Mark Zuckerberg of Meta, Greg Brockman of OpenAI, Elon Musk of xAI and Jensen Huang of Nvidia, alongside Trump’s.
It sets four layers of checks: internal monitoring of model capabilities in areas including cybersecurity, biosecurity and chemical threats, an internal team that verifies those controls, an independent external auditor or evaluator, and an independent board committee. It does not name auditors, set deadlines or require anyone to publish audit findings. The signatories say only that codifying the steps into law “may make sense” over time. Sean O’Kane noted on Equity that the signed document misspelled the United States.
The task force adds a chair, not a regulator
The Super Intelligence Force announced on Sunday 4 October names four leaders: Clayton as chair, Federal Trade Commission chairman Andrew Ferguson, Emil Michael, Under Secretary of War for Research and Engineering, and Scott Kupor, director of the Office of Personnel Management. It coordinates rather than regulates. No new rule-making power came with it.
| Date | Move | What it changes | What it leaves alone |
|---|---|---|---|
| 22 September | Rename announced at the UN | Official vocabulary | Everything else |
| 29 September | Executive Order 14434 | Federal documents and websites | The legal definition, regulations, contracts |
| 29 September | Joint Commitment on Frontier Responsibilities | Four voluntary layers of checks | Penalties, deadlines, disclosure |
| 4 October | Super Intelligence Force | A named chair and four leaders | Rule-making powers |
Read together, the AI rebrand is a communications package. The words change at once; the obligations are left for later.
Why the White House Wants an AI Rebrand Now
A government does not rename a technology it believes the public already loves. The timing of this AI rebrand tracks a long slide in public sentiment.
The polling backdrop
Pew Research Center found in June 2025 that 50% of US adults felt more concerned than excited about the growing use of AI in daily life, up from 37% in 2021. Only 10% felt more excited than concerned, down from 18%. A further 38% said they felt both equally.
The numbers have not improved since. As we reported in our analysis of Trump’s voluntary AI safeguards, a Reuters/Ipsos poll published on 22 September found 73% of Americans worry that AI companies “haven’t gone far enough to prevent AI from causing serious harm to society”. A UMass Amherst poll run by YouGov in August found 57% said Trump was handling AI poorly, and 65% opposed an AI data centre in their own community.
How US adults feel about the growing use of AI, Pew Research Center (per cent)
Each bar is drawn on a 0 to 100 scale. Concern rose by 13 points (50 minus 37) while excitement fell by 8 points (18 minus 10), so the gap between the two moved from 19 points to 40 points in four years. That widening gap is the image problem in one number.
What the TechCrunch hosts heard
On Equity, Anthony Ha pointed out how quickly the pact came together. Only a week or two earlier, he said, Trump had been arguing that “we don’t need any regulation”. Ha’s conclusion was that the agreement itself “seems not super meaningful”, but “the fact that this kind of gesture was needed seems significant”.
Korosec went further. To her, the rename was the point of the whole exercise: a way to separate the frightening thing people read about from the new, approved thing the government wants to promote. That reading makes the AI rebrand a deliberate piece of message discipline rather than a vocabulary quirk.
Mixed messages inside the administration
O’Kane added a reason for doubt. Amodei was invited to a late dinner with Trump days before the lunch, yet the Defense Department still treats Anthropic as a national security risk, and Emil Michael, now one of the four leaders of the new task force, criticises the company on X “every other day”. An AI rebrand is hard to sell to the public when the people delivering it are sending opposite signals about the companies involved.
Can a New Name Change How People Feel About a Technology?
The idea behind any AI rebrand is that words shape attitudes. That idea has real evidence behind it, but the evidence points in a less comfortable direction for the White House than it might hope.
When renaming worked: NMR became MRI
Doctors once called magnetic resonance imaging “nuclear magnetic resonance imaging”. In the 1980s the American College of Radiology’s commission on the technology recommended dropping “nuclear”, because patients associated the word with radiation, bombs and reactor accidents. The scan involves no radioactivity at all; “nuclear” referred to hydrogen nuclei. By the mid-1980s most of the field said MRI.
The rename stuck because the experience matched the new name. Patients lay in a scanner, heard loud knocking and walked out unharmed. Nothing they saw contradicted the friendlier word. That is the condition an AI rebrand would need to meet, and it is the condition that is hardest to meet for a technology that is regularly in the news for misbehaving.
When words moved opinion: global warming and climate change
The Yale Program on Climate Change Communication and George Mason University tested the two terms in 2014. Americans were 13 points more likely to call “global warming” a bad thing (76%) than “climate change” (63%). The older term produced more worry and more certainty that the problem was real.
The lesson for an AI rebrand runs both ways. Names do change emotional reactions, measurably, which is the premise of every AI rebrand. But the softer term reduced concern without changing the underlying facts, which is a communications win and a policy loss if the concern was justified.
When the word itself costs trust
Washington State University researchers tested the label directly in 2024. In experiments with more than 1,000 US adults, published in the Journal of Hospitality Marketing & Management, identical product descriptions performed worse when they mentioned “artificial intelligence”. “When AI is mentioned, it tends to lower emotional trust, which in turn decreases purchase intentions,” said lead author Mesut Cicek. The effect held across eight product and service categories and was strongest for high-risk products such as medical devices and financial services.
That finding is the strongest argument for some kind of AI rebrand. If the two letters themselves carry a trust penalty, swapping them is not irrational. The question is whether “super intelligence” carries a smaller penalty or a bigger one.
Why “super intelligence” is a strange choice for an AI rebrand
For the public, super intelligence is not the safe word. It is the frightening one. In a Future of Life Institute poll of 2,000 US adults, fielded from 29 September to 5 October 2025, 64% said superhuman AI should not be developed until it is proven safe and controllable, or should never be developed. Only 5% supported the status quo of fast, unregulated development, and 73% wanted robust regulation.
So the AI rebrand renames the everyday tools people use for email and search after the one category of the technology that polls worst. Korosec’s reading assumes “super intelligence” sounds benign. The only large poll that asked about it directly suggests the opposite.
| Case | Old term | New term | What happened to opinion |
|---|---|---|---|
| Medical imaging, 1980s | NMR imaging | MRI | Fear faded; the experience was safe |
| Climate, Yale test 2014 | Global warming | Climate change | “Bad thing” fell from 76% to 63% |
| Product labels, WSU 2024 | No label | “Artificial intelligence” | Lower emotional trust and purchase intent |
| US government, 2026 | Artificial intelligence | Super Intelligence | Untested; the new term polls worse |
What the Non-Binding Pact Does for AI's Image
A name is the cheap half of an AI rebrand. The pact was meant to be the substance. Its weakness, from a trust point of view, is not that it is voluntary. It is that it is invisible.
Trust needs evidence the public can see
Every layer in the pact that backs the AI rebrand happens behind closed doors. Internal controls, an internal team, an external auditor and a board committee all produce reports, and none of those reports has to be published. A member of the public has no way to tell whether a signatory is following the pact or ignoring it.
That matters because trust is built on observed behaviour, not stated intention. The MRI rename worked because millions of patients could check it for themselves. Nobody outside the companies can check the pact.
Disclosure beat silence in the GMO fight
The best evidence for what does work comes from food labelling. Vermont required labels on genetically engineered food from 1 July 2016 until a federal law overrode the rule. Jane Kolodinsky of the University of Vermont and Jayson Lusk of Purdue analysed the attitudes of more than 7,800 consumers between 2014 and 2017 and published the result in Science Advances in 2018.
Opposition to genetically engineered food fell by about 19% in Vermont after the labels appeared, while it kept rising in the rest of the country. Telling people more reduced fear. An AI rebrand built on a pact that tells people less is betting on the opposite effect.
Regulators are already testing the promises
The day after the signing, the FTC confirmed a broad investigation into OpenAI, Anthropic and other developers. Public promises about safety can be examined under consumer protection law, which gives the pact a route to consequences that its own text does not provide. Yet that route works through enforcement after the event, not through the reassurance an AI rebrand is meant to offer.
What US adults want for advanced AI, Future of Life Institute poll, 2025 (per cent)
Each bar is drawn on a 0 to 100 scale using the poll’s published figures. Supporters of robust regulation outnumber opponents of strong regulation by roughly six to one (73 divided by 12 is 6.1). A voluntary pact sits on the side of the 5%, not the 73%, which is why it does little for the image problem on its own.
Where the AI Rebrand Could Backfire
A rebrand that fails is worse than no rebrand, because it spends credibility. There are three ways this one could go wrong.
It attaches the scariest word to the most ordinary tools
Most people meet the technology as a chatbot, a writing assistant or a search summary. This AI rebrand calls those tools “super intelligence”, which raises expectations that they cannot meet and links them to the long-term risks that worry the public most. Every hallucinated answer or misbehaving agent now becomes a story about super intelligence failing.
The record this year makes that likely. We have covered OpenAI’s struggle to contain rogue AI activity, agents probing government websites and models gaming their own tests. An AI rebrand cannot outrun that news flow; it can only rename it.
It splits the vocabulary at home
The order binds the executive branch, not Congress, the courts, the states or companies. Statutes still say artificial intelligence. Contracts signed before the order still say it. Newsrooms, including TechCrunch, kept writing “AI” in the same articles that reported the rename. The likely result of the AI rebrand is two vocabularies, one official and one everyday, which tends to draw attention to the official term rather than replace the everyday one.
The rest of the world will not follow
The AI rebrand has no reach outside the US executive branch. The EU’s AI Act, the UK’s AI Security Institute and the international bodies that set standards all keep the old term. For multinational companies, the AI rebrand adds a translation step for American government work and changes nothing anywhere else.
What Would Actually Fix AI's Image Problem
If the goal is public trust rather than a better headline, the research points to four steps. None of them requires an AI rebrand.
Publish what the auditors find
The pact already creates the reports. Publishing a summary of each external audit, even a redacted one, would turn a private promise into something the public can check. That is the GMO lesson in practice: disclosure reduced fear where silence had fed it.
Name the auditors and give them access rules
The pact lets each company choose its own evaluator. Naming the evaluators and stating what they may inspect would answer the first question any sceptic asks, which is who checks the checkers.
Put a date on codification
The signatories say turning the steps into law “may make sense” over time. A target date would show the pact is a bridge to rules rather than a substitute for them, which is what the 73% in the FLI poll say they want.
Let people feel the benefits
MRI’s rename worked because patients experienced a safe scan. The equivalent for the AI industry is products that are reliably useful and visibly controlled, so that daily experience contradicts the scary headlines rather than confirming them.
| Driver of the image problem | What the rebrand and pact offer | What would help more |
|---|---|---|
| Fear of harm | Private internal controls | Published audit summaries |
| Doubt about who checks | An unnamed external evaluator | Named evaluators with access rules |
| Demand for regulation | “May make sense” to codify | A date for draft legislation |
| Distrust of the word “AI” | A new official name | Describing benefits, not buzzwords |
| Jobs and local data centres | Nothing specific | Local consultation and visible commitments |
What the AI Rebrand Means for UK Businesses
British firms do not have to adopt any of this vocabulary, but the AI rebrand still reaches them in three ways.
Your customers still say AI, and still distrust it
The WSU research applies on this side of the Atlantic too. Cicek’s advice to marketers was blunt: “Focus on describing the features or benefits and avoid the AI buzzwords.” For a UK business, swapping “AI” for “super intelligence” would make that problem worse, not better. Describe what a tool does for the customer and how it is checked.
US government work may need a translation layer
If you sell software or services to American federal agencies, expect the AI rebrand to reach solicitations and correspondence as “Super Intelligence” or “SI”. The order does not require vendors to change their own documents, but matching the buyer’s language in proposals is ordinary good practice. Keep a simple glossary so the two terms map cleanly in contracts.
Trust is still built locally
UK customers judge your use of AI by your data protection record, your transparency and whether a person is accountable. Publishing a plain-English AI use statement, and backing it with real controls, does more for trust than any label. Our guide to AI governance consulting covers how to turn a policy into technical controls a regulator or customer can inspect.
What to Watch Next in the AI Rebrand
The AI rebrand has three near-term tests, and each one will show whether the package has substance behind the name.
The 28 November definition deadline
The order gives the White House science adviser until 28 November to propose a new legal definition of super intelligence. A proposal that simply copies the existing definition would confirm the rename is cosmetic. One that changes what is covered would have real legal effect.
The first standards meeting
The pact promises that signatories will “meet regularly to establish standards and best practices”. No date or venue has been announced. A published standard for what the external auditor must test would be the first sign that the pact is more than a press event.
The FTC investigation and the task force’s first moves
The FTC’s investigation and the Super Intelligence Force’s first actions will show whether the government treats the pact’s promises as enforceable. If either produces public findings, that would do more for the industry’s image than the rename ever could.
AI Rebrand FAQ
What is the AI rebrand?
It is the US government’s decision to use “Super Intelligence” and “SI” in place of “artificial intelligence” and “AI”. Executive Order 14434, signed on 29 September 2026, applies it to federal documents, websites and correspondence. The legal definition has not changed yet.
Is the safety pact legally binding?
No. The Joint Commitment on Frontier Responsibilities has no penalties, deadlines or disclosure requirements. Trump called it “morally binding”. Consumer protection law can still examine public safety promises, and the FTC opened an investigation the day after the signing.
Who signed the pact?
Sundar Pichai of Google, Dario Amodei of Anthropic, Mark Zuckerberg of Meta, Greg Brockman of OpenAI, Elon Musk of xAI and Jensen Huang of Nvidia, alongside President Trump. Microsoft and Amazon, which signed the 2023 voluntary commitments, are not on it.
Will the AI rebrand change public opinion?
Names can shift emotional reactions, as the climate and MRI cases show. But the public fears super intelligence more than ordinary AI, according to the Future of Life Institute’s polling, and research on GMO labels suggests disclosure builds more trust than new wording.
Does the AI rebrand apply in the UK?
No. It binds only the US executive branch. UK law, the EU AI Act and international standards bodies continue to use “artificial intelligence”. UK firms that sell to US federal agencies may see the new term in solicitations.
What would improve trust in AI?
Published audit summaries, named evaluators with clear access rules, a date for turning voluntary commitments into law, and products whose everyday reliability matches the claims made for them.
References and Further Reading
Can ‘super intelligence’ and a non-binding safety pact solve AI’s image problem? (TechCrunch)
Don’t say AI: Trump’s answer to AI’s image problem (Computerworld)
Key findings about how Americans view artificial intelligence (Pew Research Center)
Americans want regulation or prohibition of superhuman AI (Future of Life Institute)
Labeling decreases opposition to genetically engineered food (Purdue University)
Magnetic resonance imaging: history of the name (Wikipedia)
Trump and major AI executives sign “morally binding” voluntary controls (CBS News)
More AI coverage: explore Progressive Robot's AI Models, Tools & Releases hub — hands-on reviews, setup guides and benchmarks in one place.
