Should tech labels depend on consensus or decree?
On 29 September 2026 Donald Trump issued an executive order directing US federal agencies to use Super Intelligence in place of Artificial Intelligence, defining the new term to cover the same technologies as the old one and calling those who refused to comply “THE ENEMY”. Since its first use in this context in 1955 the term has been in and out of fashion, surviving two ‘AI winters’ when both concept and name were out of favour among specialists. Its power today derives from its endorsement by all the major stakeholders: tech bros, broligarchs, government, commerce and media, hence also the general public. The English term has been translated directly into other major languages. Ten days later I spoke to tech journalist Chris Stokel-Walker about the issues surrounding the attempted rebranding.
Chris’s article for Fast Company is here…
https://www.fastcompany.com/91621442/its-not-ai-and-its-not-superintelligence-so-what-is-it
Before the article appeared, I had asked AI engines Chat GPT and Google Gemini which of the two terms they preferred. These were their responses in their own words…
“I would choose AI, rather than SI, but the distinction is more interesting than it might at first appear. There are really two questions here: which term is more useful, and what assumptions each term encourages us to make about the technology. ‘AI’ or Artificial Intelligence is the established, relatively neutral term. It describes a broad field of technologies, from relatively simple pattern-recognition systems to sophisticated language models such as myself. It makes no necessary claim that a system is conscious, wise, autonomous or superior to human beings. ‘SI’ – Superintelligence – is a much stronger term. It suggests intelligence that surpasses human capabilities, perhaps across most or all intellectual domains. It is not simply another name for AI; it describes a hypothetical or prospective level of capability. Using it can therefore make a prediction about the future sound like a description of an existing reality.”
“If you are looking at the natural stepping stones of computer science, the missing link between today’s regular AI and actual Super Intelligence (SI / ASI) is AGI (Artificial General Intelligence). AI: Narrow systems designed for specific tasks (like ChatGPT or AlphaFold). AGI: A hypothetical machine that possesses human-level reasoning, learning, and adaptability across any domain while SI here implies an intellect that drastically surpasses human capabilities in every cognitive field. An alternative might be Synthetic Intelligence (SI). Interestingly, before “SI” was adopted by the U.S. government to mean “Super Intelligence”, the abbreviation SI already stood for Synthetic Intelligence in tech circles. If you want a third term that avoids the baggage of “artificiality,” you can look at Cognitive Computing: Systems that mimic the human brain’s thought processes using pattern recognition and natural language processing. This can subsume Embodied Cognition / Embodied AI: AI that interacts physically with the real world through sensors and robotics to learn like a human child, rather than just reading data on a screen. If you are looking at the broader philosophical “pillars of intelligence,” the ultimate third alternative is NI (Natural / Biological Intelligence). This represents human and animal cognition—the benchmark that AI attempts to mimic and SI aims to surpass.”
