US AI Safety Institute renamed Center for AI Standards and Innovation

Commerce Secretary Howard Lutnick announced that the US AI Safety Institute, the NIST body created under the Biden administration to evaluate frontier model safety, would be renamed the US Center for AI Standards and Innovation (CAISI). The change was framed as more than cosmetic: the department described the institute’s prior work as having strayed into bias, misinformation and other contested social questions, and said the renamed body would instead pursue what it called pro-innovation, pro-science work aimed at keeping the US ahead of foreign competitors.

The renamed body’s mandate shifted toward national-security-relevant risks — cybersecurity, biological and chemical weapons potential, and the capabilities of foreign, particularly Chinese, models — and toward pushing back against what the administration characterised as burdensome AI regulation by other governments and international bodies, rather than the broader model-evaluation and public-interest safety work the institute had pursued under its previous mandate and name.

The change followed the Trump administration’s earlier rescission of Biden’s 2023 AI executive order, and confirmed a pattern set by that action: the federal government’s institutional machinery for AI oversight, built in 2023 and 2024 around a “safety” framing that treated bias, misinformation and societal harm as central concerns, was being repurposed under a framing centred on competitiveness and national security instead. The rebrand left the institute’s technical evaluation capacity largely intact while narrowing what counted as a risk worth evaluating.