You Can Call it SI. It’s still AI

Image generated with AI

So, apparently, Artificial Intelligence is no longer Artificial Intelligence. At least not if you work for the executive branch of the U.S. federal government.

On September 29, President Donald Trump signed an executive order titled Inaugurating the Era of Super Intelligence. The order establishes a policy directing executive departments and agencies, to the maximum extent permitted by law, to use “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI” in official communications and other non-statutory documents.

The accompanying White House fact sheet presents the change as recognition that today's technologies have moved beyond what the term "artificial intelligence" originally represented.

However, there is one particularly interesting detail:

For purposes of implementing the order, SI initially means the technologies already encompassed by the existing statutory definition of artificial intelligence. The order then directs the Assistant to the President for Science and Technology to propose a federal definition of Super Intelligence, or SI, within 60 days.

At least for now, much of what changed is the terminology, so, consequently, I will keep calling it AI.

Don’t get me wrong, it’s not because AI is a perfect term; it isn't. But because changing the name doesn't change the technology.

AI Is Already an Imperfect Name

Let's acknowledge something first: Artificial Intelligence has always been an awkward label. The field has spent decades debating what exactly “intelligence” means. Technologies placed under the AI umbrella have also changed dramatically, from expert systems and machine learning to deep learning, generative AI, foundation models, and now increasingly autonomous agents.

Today's frontier systems would certainly have surprised many researchers working when the field was established in the 1950s, but, in my view, technological progress alone doesn't make today's systems “superintelligent.”

This term/adjective already carries substantial conceptual baggage.

Within AI research and public discussion, superintelligence has commonly been associated with hypothetical systems whose intellectual capabilities substantially exceed human abilities across broad domains.

This is very different from saying today's systems can outperform humans at tasks. Current AI can write software, analyze enormous datasets, generate images, summarize documents, identify patterns, and increasingly coordinate tools and workflows. Impressive? Absolutely.

Superintelligent? I think it requires a much more careful discussion, not a political stance.

Naming Something Doesn't Make It So

And this is where I become uncomfortable with politically driven technological terminology.

Obviously, governments have the authority to establish terminology for their own programs and administrative operations. They can define categories, establish policies, fund research, regulate industries, and create national technology strategies.

But scientific terminology should ideally emerge from evidence, research, technical consensus, and sustained debate, even if it is still imperfect or somehow inaccurate.

So, a political declaration cannot change the underlying capabilities of a neural network. If tomorrow we rename quantum computing “Ultra Computing,” quantum computers will not suddenly become more powerful. Likewise, replacing AI with SI doesn't make models more intelligent.

Words matter because they shape expectations, and "super intelligence" is not a neutral term; it implies superiority, and it suggests that we have crossed an important technological threshold, perhaps someday we will, but claiming the destination before demonstrating that we have arrived risks replacing technical description with aspiration.

Keep Politics Away from the Definition of Technology

Now, to clarify, this isn't really an argument about President Trump, Democrats, Republicans, or any administration; ultimately, another government could make a similarly problematic decision in the opposite direction, and the principle should remain the same.

Science and technology need room to describe reality independently of political messaging.

Sure, governments inevitably influence technology; they fund research, procure systems, establish standards, regulate markets, and determine national priorities. This relationship is unavoidable, even often necessary.

But there is an important difference between establishing technology policy and redefining scientific terminology to reinforce a political or economic narrative. Because once terminology becomes political branding, it can become unstable.

One administration calls something AI; another calls it SI. A future administration could invent another term.

Meanwhile, researchers, engineers, businesses, universities, standards organizations, and international institutions still need a shared vocabulary with which to communicate. Technical language works best when it helps us understand reality rather than tell us how impressed we should be by it.

The Industry Doesn't Need More Hype

I think there is also a practical industry problem. AI already suffers from “terminology inflation." First, everything became AI, then everything became generative AI, then copilots, then agents, now agentic AI.

And while artificial general intelligence (AGI) remains somewhere on the horizon, now we have SI.

Second, enterprise technology buyers already struggle to distinguish actual capabilities from product positioning; introducing increasingly dramatic terminology before clear technical distinctions exist doesn't make those decisions easier. it makes them harder.

A CIO evaluating an AI platform doesn't need a more exciting acronym; he needs to know what the system can do, what data it requires, how accurate it is, how much it costs, how securely it operates, where it fails, and of course whether it produces measurable business value.

Simply calling the technology Super Intelligence answers none of those questions.

AI Is Good Enough for Us

So, for D of Things, I intend to continue using Artificial Intelligence and AI unless I'm specifically discussing the U.S. government's SI terminology or a technically defined form of superintelligence.

Don’t get me wrong, there is nothing rebellious about that decision. It is simply that I consider AI to be imperfect but still more precise.

AI remains the broadly understood umbrella for a family of technologies that includes machine learning, deep learning, generative systems, computer vision, language models, and increasingly agentic architectures.

If genuine superintelligence emerges someday, I'll happily discuss what we should call it.

But first, let's demonstrate that it exists, and then we can worry about the acronym.

Meanwhile, We're Ignoring Harder Questions

And perhaps this is what bothers me the most about the naming debate. While we discuss whether two letters should become two different letters, considerably more important AI questions remain unresolved:

• Who is accountable when an autonomous agent makes a consequential mistake?
• How should organizations govern increasingly autonomous systems?
• What ethical boundaries should constrain their use?
• How do we audit decisions made through combinations of models, data, tools, and agents?
• Who bears responsibility when automated decisions cause economic, professional, or physical harm?
• How much transparency should organizations provide when AI participates in decisions affecting people?
• And how do we ensure that governance evolves alongside capability rather than several years behind it?

These questions aren't as exciting as declaring the beginning of a new technological era, but, unfortunately, they are considerably more important.

Reality Before Aspiration

Technology doesn't become transformative because we give it a transformative name. It becomes transformative when its capabilities materially change what people and organizations can accomplish, and AI is already doing that. There’s no need to exaggerate it.

Perhaps someday we will genuinely enter an era of superintelligence. If that happens, the evidence should make the terminology obvious.

Until then, I prefer language that describes where the technology is or has been representing through its evolution, not where governments, vendors, investors, or enthusiasts want us to believe it is going.

So yes, Washington can call it SI; for now, I'll keep calling it AI.

And rather than spending too much energy debating the acronym, I would rather focus on something much harder:

Making sure that whatever we call these increasingly powerful systems, we build the accountability, governance, ethics, transparency, and human responsibility necessary to use them wisely.

Because ultimately, history probably won't care what acronym we chose. It will care what we did with the technology behind it.

Top of Form

Bottom of Form

But what do you think?

Please feel free to share your perspective.

Until next time,

Jorge Garcia