119-HR-5784 Corporate Impact Analysis

Summary

• Scope: Directs SBA to stand up public, non‑preferential AI learning modules on its existing online platform, consulting NIST and an advisory working group; no new funds authorized. [1]U.S. Small Business Administration — SBA Learning Platform | U.S. Small Busines…[2]National Institute of Standards and Technology — Artificial Intelligence Risk M…[3]National Institute of Standards and Technology — AI RMF Generative AI Profile |…
• Economic signal: Education and standardized risk practices should lower search, vendor‑selection, and compliance costs for small firms; adoption impact likely modest but positive from a low baseline. [4]U.S. Census Bureau — Tracking Firm Use of AI in Real Time: A Snapshot from the…
• Operational upside: Peer‑reviewed evidence shows double‑digit productivity gains from GenAI assistants in support workflows, especially for less‑experienced workers. [5]Stanford Graduate School of Business / QJE — Generative AI at Work (QJE 2025) |…
• Externalities: Any added AI uptake carries data‑center energy and local water‑use implications; global power demand from data centers is projected to more than double by 2030. [6]International Energy Agency — Energy and AI – Executive summary | IEA
• Governance risk: Exempting the advisory working group from FACA can reduce meeting openness and disclosure norms unless SBA adopts compensating transparency practices. [7]U.S. General Services Administration — When is FACA applicable? | GSA[8]U.S. Government Accountability Office — Federal Advisory Committees: Actions Ne…

Economic Effects

Impacts framed on cost, compliance, and competitive dynamics for small firms and AI vendors.

• Lower evaluation and compliance costs: Modules anchored to NIST’s voluntary AI RMF and GenAI Profile can give small firms standardized checklists (risk, privacy, vendor management), reducing reliance on paid consultants and improving vendor comparability. [2]National Institute of Standards and Technology — Artificial Intelligence Risk M…[3]National Institute of Standards and Technology — AI RMF Generative AI Profile |…
• Adoption from a low base: Census BTOS real‑time estimates show AI use among U.S. businesses rose from ~3.7% to ~5.4% (Sep 2023–Feb 2024) with ~6.6% expected by early Fall 2024—implying headroom for gains if evaluation frictions fall. [4]U.S. Census Bureau — Tracking Firm Use of AI in Real Time: A Snapshot from the…
• Productivity potential: A QJE‑published study finds ~15% average output gains for customer‑support agents using GenAI, with the largest improvements among novices—suggesting outsized benefits for thinly staffed small firms. [5]Stanford Graduate School of Business / QJE — Generative AI at Work (QJE 2025) |…
• Market access: SBA’s nationwide resource‑partner network (SBDCs, SCORE, WBCs, VBOCs) reaches over a million entrepreneurs annually, providing a ready distribution channel for the modules and potential lead‑gen for vetted AI vendors that meet neutrality rules. [9]U.S. Small Business Administration — Resource Partners | U.S. Small Business Ad…[10]U.S. Small Business Administration — SBA Resource Partners Honored at National…
• Labor and reorganization: BTOS analysis reports few firms cutting headcount due to AI; instead, users report staff training, new workflows, and cloud purchases—costs that training can help plan and sequence. [4]U.S. Census Bureau — Tracking Firm Use of AI in Real Time: A Snapshot from the…
• Budget and implementation: “No additional amounts authorized” implies SBA must reallocate within existing appropriations—limiting scale but avoiding new fees or mandates on businesses (execution risk if resourcing is thin).

Social Effects

Distributional and compliance implications for communities and workforce.

• Access and inclusion: Delivery via SBA partners can reach women‑, veteran‑, and rural‑owned firms through WBCs and VBOCs, improving digital‑skills equity. [9]U.S. Small Business Administration — Resource Partners | U.S. Small Business Ad…
• Civil‑rights compliance: EEOC has active guidance on algorithmic hiring and adverse impact; awareness training can lower litigation risk for small employers adopting AI in hiring. [11]U.S. Equal Employment Opportunity Commission — EEOC 2023 Annual Performance Rep…
• Legal exposure signal: Courts are allowing novel AI‑bias suits against HR‑tech platforms to proceed, underlining the value of modules on human oversight and vendor due diligence. [12]Reuters — Workday must face novel bias lawsuit over AI screening software
• Consumer‑protection posture: FTC guidance flags deceptive AI marketing and AI‑enabled impersonation risks; training can help firms avoid unfair or deceptive practices in advertising and customer engagement. [13]Federal Trade Commission — Chatbots, deepfakes, and voice clones: AI deception…[14]Federal Trade Commission — FTC Proposes New Protections to Combat AI Impersonat…[15]Federal Trade Commission — FTC Final Rule Extends Telemarketing Fraud Protectio…

Environmental Effects

Indirect impacts arise if the program raises AI utilization among small firms.

• Electricity demand: IEA projects data‑center power use to more than double to ~945 TWh by 2030, with AI as a primary driver; U.S. accounts for the largest share of projected growth. [6]International Energy Agency — Energy and AI – Executive summary | IEA
• U.S. context snapshot: Analysts estimate U.S. data centers used ~183 TWh in 2024; concentration in certain states can strain local grids—important for siting and tariff risk. [16]Pew Research Center — US data centers’ energy use amid the AI boom | Pew Resear…
• Water footprint: Cooling and power generation make AI workloads water‑intensive in some regions; indirect (off‑site) water use typically dominates direct facility use. [17]IEEE Spectrum — The Real Story on AI Water Usage at Data Centers | IEEE Spectrum
• Mitigation levers: Efficiency guidance (model choice, batching, caching) and provider‑level engineering can reduce intensity; awareness modules could incorporate such practices to temper incremental load. [17]IEEE Spectrum — The Real Story on AI Water Usage at Data Centers | IEEE Spectrum

Temporal Analysis

Short‑term versus long‑term effects, assuming enactment.

Unintended Consequences

• De facto obligations: Even with a “voluntary” posture, widely used modules aligned to NIST RMF could become de facto expectations from customers or insurers—raising soft compliance burdens for late adopters. [2]National Institute of Standards and Technology — Artificial Intelligence Risk M…
• Regulatory spillovers: As state AGs increasingly police AI under existing privacy/consumer‑protection laws, firms following SBA guidance may still need jurisdiction‑specific controls. [18]Reuters — State AGs fill the AI regulatory void
• Over‑reliance on vendor content: If neutrality guardrails aren’t enforced, materials could privilege incumbent providers; SBA should curate against marketing claims flagged in FTC guidance. [13]Federal Trade Commission — Chatbots, deepfakes, and voice clones: AI deception…

Assessment

Overall stance reflects net business impact, regulatory risk, and execution feasibility.

Favorable (modest, execution‑dependent). The bill imposes no new mandates or fees on businesses, uses existing SBA infrastructure, aligns with recognized risk frameworks (lowering compliance friction), and targets documented capability gaps that limit small‑firm AI uptake. Risks—especially FACA‑exempt governance and environmental externalities from increased AI use—are real but manageable with voluntary transparency, careful curation, and inclusion of efficiency practices in coursework. [1]U.S. Small Business Administration — SBA Learning Platform | U.S. Small Busines…[2]National Institute of Standards and Technology — Artificial Intelligence Risk M…[4]U.S. Census Bureau — Tracking Firm Use of AI in Real Time: A Snapshot from the…[6]International Energy Agency — Energy and AI – Executive summary | IEA[8]U.S. Government Accountability Office — Federal Advisory Committees: Actions Ne…

Key Metrics

Sourcing Notes

Core figures and risk frameworks draw on U.S. government statistics, peer‑reviewed research, and major institutions (NIST, IEA).

• Where the analysis infers impacts (e.g., potential de facto standards), we identify the basis (voluntary frameworks widely referenced by industry). [2]National Institute of Standards and Technology — Artificial Intelligence Risk M…
• Environmental projections reflect sector‑level trends; local effects will vary by grid mix and water scarcity. [6]International Energy Agency — Energy and AI – Executive summary | IEA[17]IEEE Spectrum — The Real Story on AI Water Usage at Data Centers | IEEE Spectrum