Super Intelligence (SI)

Super Intelligence, or SI, now appears in political debate about artificial intelligence. In technical discussions, superintelligence generally means intelligence beyond human capabilities across a broad range of tasks. These uses need careful separation: a government's choice of terminology does not, by itself, establish a scientific breakthrough.

On September 29, 2026, President Donald Trump directed U.S. executive agencies to use “Super Intelligence” in official communications and certain documents, within legal limits. For current implementation, the order retains the existing statutory definition of AI. It also calls for proposed legislation on that definition to be submitted to the President within 60 days.

This hub begins with our analysis of that order and a separate voluntary industry safety accord. The questions concern what changes in practice: how agencies implement the wording, whether a proposed definition becomes public, and what evidence companies release about their safety commitments.

Our coverage will examine policy documents, independently assessed capabilities and responsibility for deployment decisions. When companies or officials describe a system as superintelligent, the follow-up should be specific: what can it do, how was it tested, and who can verify the result? We distinguish documented findings, attributed claims and unresolved questions.

The SI story connects government terminology, technical capabilities and company oversight. The September 2026 U.S. order changes executive-branch language while retaining the existing AI definition for implementation and requesting a legislative proposal. A separate voluntary industry accord promises safety reviews but leaves questions about timing and public disclosure open. Evaluating what changes requires examining agency instructions, proposed legislation, evaluation methods and published findings.