The Foundation for American Innovation (FAI), a Washington technology policy think tank, has renamed its AI Policy program Super Intelligence Policy. In an announcement published on November 2, the group said the pace of progress in frontier models has turned true Super Intelligence from a science-fiction idea into a pressing question of national readiness.
The team behind the program will now be known as FAI's SI Team. Its stated job is to prepare lawmakers and officials for what the foundation calls the Super Intelligence era.
Why FAI changed the name
FAI points to how quickly the field has moved since its AI program started in April 2023. Back then, chatbots such as Claude and ChatGPT stumbled over simple arithmetic. According to FAI, frontier models now produce math results worthy of a Fields Medal within an afternoon, for a few hundred dollars, and progress is compounding with no sign of a slowdown.
The foundation also draws a line between the two terms. In its view, "artificial intelligence" is a catch-all that stretches from chess programs to spam filters. Super Intelligence (SI), by contrast, describes persistent, goal-driven agents that beat human experts in several areas at once and may soon outperform people in every kind of thinking.
Already Super Intelligent?
FAI goes further than most groups by arguing that current frontier models are already Super Intelligent in a commonsense sense. It lists what they can do: work together in large teams on one problem for weeks without tiring, operate any software or tool, and find security flaws that sat unnoticed for decades. The foundation adds that these systems are starting to make real new contributions to science.
Yet the technology underneath has changed little, FAI notes. These systems are still deep neural networks trained on vast amounts of data. The group cites physicist Philip W. Anderson, the Nobel laureate whose 1972 essay More Is Different argued that scale can create entirely new behavior.
"Perhaps the most important policy question of our time."
— FAI
A new policy field
For FAI, that is why SI policy cannot be treated as AI policy with a new label. Handling it as a follow-on to the past decade's fights over social media algorithms would be a category error, the foundation says, and the era calls for fresh laws and concepts.
FAI says it will keep covering the wider range of AI and technology issues, but will avoid lumping ordinary algorithms together with frontier Super Intelligence. It urges the current administration, and any future one, to keep the same distinction for the sake of clarity.
The announcement closes on a lighter note. Comparing the switch to the Gulf of America and Twitter becoming X, FAI is releasing an "AI → SI" Chrome extension that automatically replaces "AI" with "SI" and "artificial intelligence" with "super intelligence" across the web.
Why it matters
FAI is one of the first policy organizations to rebuild a whole program around the Super Intelligence label, following the White House order on the term. Its move shows the vocabulary spreading from government and big tech into the think tanks that shape legislation. The next thing to watch is whether other policy groups and lawmakers adopt SI as a separate field, as FAI predicts they will.