What happened
Nuclear Regulatory Commission Chief Data Officer and Deputy Chief AI Officer Basia Sall said on June 25, 2026 that AI has already shortened NRC licensing reviews. Sall is quoted in Nextgov/FCW: "For example, one type of licensing would take four years. We said we're going to get it down to 18 months. We just finished that first round of that licensing in nine months." She attributed the compression to AI tooling used alongside a May 2025 Trump executive order that set an 18-month target, per Nextgov/FCW.
How the NRC is using AI
Per Nextgov/FCW, the NRC has deployed AI for drafting documents and reviewing precedent from prior decisions. The agency has engaged industry partners that curated NRC public data into structured datasets; Sall said this yields clearer applications and fewer follow-up questions during review. The NRC has also leveraged GSA AI offerings and partnered with GSA on an AI maturity assessment and strategic planning process, FedScoop reports.
Broader context
The Department of Energy and national laboratories have been separately collaborating with the NRC on AI-assisted review workflows, with DOE publishing plans to use AI to reduce reactor licensing timelines, per DOE and the American Nuclear Society. The NRC has an active AI strategy page (nrc.gov/ai) and has issued strategic plans for reviewing AI in nuclear applications.
Why it matters for practitioners
The NRC case illustrates several production requirements for document-AI in high-assurance regulatory contexts: structured retrieval from large precedent sets, secure data-sharing arrangements for industry-curated datasets, and submission quality improvement as a side effect of better tooling on the applicant side. The 4-year-to-9-month result is a single data point from one licensing category -- independent audits of model choices and human-in-the-loop controls have not yet been published.
What to watch
Whether the NRC publishes technical assessments or audit trails documenting model choices, data governance, and human review steps; whether other regulators cite NRC experience when procuring similar tools; and progress on the DOE-NRC collaboration for reactor licensing.
Key Points
- 1NRC CDO Basia Sall reports AI compressed one licensing type from four years to nine months -- exceeding the 18-month executive order target set in May 2025.
- 2The NRC uses AI for document drafting and precedent review; industry partners curating NRC public data improved submission quality and reduced back-and-forth.
- 3DOE and national labs are separately collaborating with the NRC on AI-assisted review workflows, suggesting this is becoming a multi-agency effort.
Scoring Rationale
A concrete, on-the-record government AI deployment result -- one licensing category cut from four years to nine months -- with corroborating DOE/national lab collaboration. Solid niche-but-relevant story for practitioners building document-AI in regulated contexts; limited to a single agency's reported outcome without independent audit.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems