U.S. President Donald Trump is trying to rename the "Artificial Intelligence" (AI) that has been in use for decades as "Superintelligence" (SI). In September this year, he first accused the word “artificial” of implying falsehoods in public speeches, and then signed an executive order requiring federal agencies to replace “AI” with “SI” in official communications. More bluntly, he also said on social media platforms that anyone who continues to use “Artificial Intelligence” will be regarded as a “enemy” by the White House.
This pressure is seeping into Silicon Valley. Musk took the lead in expressing his loyalty by renaming SpaceXAI to SpaceXSI; Benioff of Salesforce followed suit, changing AIForce to SIForce; Jensen Huang of NVIDIA not only publicly supported “superintelligence research” but also announced a $1 billion investment in related scientific work; Nadella of Microsoft also repeatedly used new terms in articles discussing AI security, and those who attended the meeting included Zuckerberg, Pichai, and Bezos.
The concepts are flattened, making it difficult to resolve the disputes.
Critics point out that this rebranding is more of a political tactic than a technical clarification. AI encompasses speech recognition, image generation, and large language models—with vastly different capabilities. “Superintelligence” in academia refers to systems that completely surpass human cognition. Nick Bostrom’s 2014 book specifically discusses the risks associated with this concept. Mixing the two might dilute the significance of genuine breakthroughs, and it also puts companies in a dilemma: they may get attention, but it becomes difficult for them to highlight future upgrades through naming. Between catering to policies, competing for business opportunities, and accurately describing technology, the balance in Silicon Valley is just beginning. After all, it’s easy to change a name, but the boundaries of the technology itself don’t move accordingly.