President Donald Trump's administration is retiring the term "artificial intelligence" in official government documents, adopting "super intelligence" in its place. The rebranding was announced at a technology summit on September 29. The terminology shift arrives alongside an oversight approach built on voluntary participation rather than binding mandates.

Executive Order 14409 and the voluntary review framework

The policy's foundation was laid on June 2, when Executive Order 14409 established a voluntary review framework covering what the administration terms "covered frontier models." Under the arrangement, AI developers may submit their most advanced systems for cybersecurity reviews lasting up to 30 days before public release.

The framework's defining feature is its voluntary nature. The administration has been careful to distinguish the process from anything resembling a licensing regime or pre-clearance requirement: companies may participate if they choose, and there is no penalty for skipping the process entirely. The practical upshot is that frontier systems can reach public release without any federal review, and the decision — along with any preparation it requires — rests entirely with the companies building the models.

Trump has framed the approach as a middle path between reckless permissiveness and heavy-handed regulation. Rather than preventing misuse through gatekeeping, the existing legal apparatus — particularly the Department of Justice — is positioned as the backstop for catching misuse after it happens.

Rejection of international governance

The administration's stance on international AI governance was made unmistakable on September 22, when Trump declared in his speech at the UN General Assembly the United States rejects what he called "globalist control schemes" for super intelligence. Taken together, the domestic framework and the UN speech define the administration's posture: cooperation at home is voluntary, and international control frameworks are rejected.

What sparked the cybersecurity focus

The voluntary review framework did not emerge from a vacuum. Internal debates within the administration were reportedly catalyzed by concerns over advanced models such as Anthropic's Mythos series, which demonstrated remarkable capabilities in the cyber domain.

The framework's scope is deliberately narrow, focusing on cybersecurity rather than broader safety concerns. That limited focus is intentional: it keeps the review process manageable and avoids the sprawling regulatory apparatus that a more comprehensive safety framework would require. It also tracks with the concerns that reportedly set the process in motion. What to watch from here is participation: because abstention carries no penalty, how many frontier-model developers opt into the 30-day reviews will show how much practical weight the voluntary framework actually carries.