# Craft brief — Artificial Intelligence Risk Evaluation Act of 2025 (S. 2938 IS)

**Print:** 119th Congress · S. 2938 · Introduced in Senate (IS) · 2025-09-29  
**Sponsors:** Hawley, Blumenthal  
**Source:** GovInfo HTML → `prep.gpo_bill_html` → KBMill manufacture (Small, host-only)  
**Brick id:** `S2938_AI_Risk_Evaluation_Act_119IS` · public slug `s-2938-ai-risk-evaluation`

## Purpose
Residual-honest package of **what this instrument says** — Advanced Artificial Intelligence Evaluation Program at DOE, definitions (including compute thresholds), duties, and related provisions — so a local or frontier model can quote the print instead of a news recap.

## What it can quote
- Short title and sense of Congress / purposes  
- Definitions in SEC. 3 (advanced AI system, adverse AI incident, covered developer, etc.)  
- Program establishment and activities as written in this IS print  

## What it must NOT do
- Not legal advice. Not a position for or against the bill.  
- Do not treat this IS print as enacted law.  
- Do not invent later substitutes, House companions, or floor amendments not in this brick.  
- Forecast / extinction rhetoric is **not** in this instrument — say so if asked.

## Residual / trust
- Extraction quality on this HTML path: strong (quality 1.0 on residual board).  
- Soft ship warnings only (no hard fail).  
- U.S. Government work. Milled at KBMill. Not a Congress chatbot.

## Adjacent (shelf, when live)
- FRONTIER Act (House) — different coverage / audit design  
- Secure A.I. Development Act (S. 5061) — incidents / tracking  
- CREATE AI (S. 4441) — NAIRR research resource