Essay

You Can't Executive Order Your Way to Superintelligence

Published October 2026 · Back to writing
AI GovernanceTechnology LeadershipDigital Trust

Artificial intelligence has apparently been promoted. No breakthrough required. Just a presidential signature.

Apparently, artificial intelligence has been promoted.

Not through a scientific breakthrough. Not through an extraordinary advance in reasoning, autonomy or capability.

Through an executive order.

On 29 September 2026, US President Donald Trump signed Executive Order 14434, directing federal agencies to replace the term Artificial Intelligence (AI) with Super Intelligence (SI).

Same technology. Bigger letters.

And if the technology industry is prepared to play along without questioning the distinction, perhaps the intelligence we should be worried about isn't artificial.

Words actually mean things

Artificial intelligence is an established field of computer science encompassing technologies that perform tasks associated with human intelligence.

Superintelligence means something considerably different.

In AI research, it describes a hypothetical level of machine intelligence that substantially exceeds human cognitive capabilities across a broad range of domains.

It isn't a more flattering name for ChatGPT.

It isn't a marketing upgrade.

And it certainly isn't something a president can declare into existence.

Yet the executive order manages a remarkable piece of linguistic gymnastics. It directs federal agencies to adopt the new terminology while defining Super Intelligence using the existing statutory definition of artificial intelligence.

The name has changed. The definition hasn't.

Imagine solving cybersecurity by renaming vulnerabilities resilience opportunities. Or eliminating technical debt by declaring it heritage infrastructure.

Unfortunately, reality is less accommodating than a communications strategy.

When branding becomes policy

Political leaders are entitled to articulate a vision for technological progress.

They are not entitled to expect scientific terminology to bend to their branding preferences without challenge.

And this isn't staying in Washington.

In October, Reuters reported that US Treasury officials were pressing international partners to adopt the term Super Intelligence in discussions involving the International Monetary Fund.

International institutions are being asked to accommodate a political administration's preferred terminology for an established field of technology.

This isn't technological leadership.

It's semantic theatre.

And the consequences extend beyond an unfortunate acronym.

Technical language provides a shared foundation for research, procurement, regulation and assurance. When terminology is repurposed for political messaging, the distinction between demonstrated capability and aspirational promise becomes less clear.

That distinction matters when organisations are making consequential decisions about systems they may not fully understand.

The vendor meeting test

I've spent more than two decades working in technology. I've seen enough industry terminology come and go to recognise when language is doing more work than the technology itself.

Some terms describe genuine developments. Others have been stretched beyond usefulness by marketing departments.

Now we have Super Intelligence.

If a vendor walked into a meeting tomorrow and proudly announced that their product was powered by SI, my first question would be simple:

What does that mean that AI doesn't?

Not what does it promise.

Not what does the brochure say.

What measurable technical distinction justifies the claim?

There is a difference between adopting terminology to satisfy a government requirement and using that terminology to imply a technological advancement that hasn't been demonstrated.

The former may be a commercial necessity.

The latter is a credibility problem.

I don't need suppliers to embrace the fashionable language of the day. I need them to explain what their systems do, where they fail, what authority they exercise and how their claims can be independently assessed.

If they cannot answer those questions, the meeting has a much bigger problem than terminology.

Intelligence is not accountability

We are deploying increasingly capable AI systems into environments where their outputs influence real decisions, services and people.

These systems do not become trustworthy because we describe them as intelligent.

And certainly not because we describe them as super intelligent.

Capability and accountability are different things.

A system might produce extraordinary results while remaining opaque about how it arrived there. It might execute complex workflows while operating beyond meaningful oversight.

It can be technically impressive and operationally irresponsible at precisely the same time.

That is why I keep returning to a principle that applies well beyond AI:

Digital trust should be observable, governable and explainable.

We need evidence of capability, clarity of authority and meaningful oversight.

A new acronym delivers none of those things.

The industry should know better

What troubles me most isn't that a politician has decided to rename a technology.

It's whether technology professionals will challenge the terminology when it obscures rather than clarifies.

Professional credibility requires the ability to distinguish technical reality from political messaging, even when that messaging comes from an influential government or an important customer.

That isn't hostility towards innovation.

It's taking innovation seriously enough to describe it accurately.

I am genuinely excited about what artificial intelligence can make possible. Its potential deserves serious attention.

But enthusiasm is not a substitute for intellectual discipline.

And scepticism is not opposition to progress.

Call it what it is

If we eventually develop systems that genuinely meet a defensible definition of superintelligence, I'll gladly acknowledge the achievement.

Until then, I'll keep calling artificial intelligence what it is.

Artificial intelligence.

Because words shape expectations. Expectations shape decisions. And decisions carry consequences.

You cannot executive order your way to superintelligence.

You can only executive order everyone to pretend.


Related: AI does not create a new domain of trust