119-HR-5764 Investigative Journalist Impact Analysis

Summary

What it does. The bill would amend 15 U.S.C. 648 to require SBDCs to provide information, guidance, training, and outreach on AI use in operations, while incorporating the statutory definition of “artificial intelligence” from 15 U.S.C. 9401. It does not authorize additional funding, so execution would depend on existing SBDC capacity. [3]Legal Information Institute (Cornell) — 15 U.S. Code § 648 - Small business dev…[1]Legal Information Institute (Cornell) — 15 U.S. Code § 9401 - Definitions (Nati…

Bottom line. If implemented with vendor‑neutral, risk‑aware guidance (e.g., mapped to NIST AI RMF), SBDC‑delivered AI assistance could improve task‑level productivity for many small firms but will also push new obligations around data security, IP, and truthful marketing; impacts will vary sharply by firm size, sector, and local infrastructure constraints. [4]NIST — NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0)[5]U.S. Census Bureau — Tracking Firm Use of AI in Real Time: A Snapshot from the…

Economic Effects

Evidence-centered estimates; not advocacy.

• Productivity and revenue potential. Field and experimental studies indicate material task‑level gains from GenAI (e.g., +14% productivity for support agents; large benefits for lower‑skilled workers; ~40% faster completion in professional writing tasks), suggesting upside for service‑heavy SMBs if SBDC training targets concrete workflows. [6]National Bureau of Economic Research — Generative AI at Work (NBER Working Pape…[7]Stanford University — Experimental evidence on the productivity effects of gene…
• Adoption reality check. High‑level surveys aside, official high‑frequency data show low but rising AI adoption among US firms (≈3.7% → 5.4% in 2023–Feb 2024; approaching ~10% by mid‑2025), with larger firms leading—implying many SBDC clients start from near‑zero and need basics (use‑case scoping, data hygiene, change management). [5]U.S. Census Bureau — Tracking Firm Use of AI in Real Time: A Snapshot from the…[8]U.S. Census Bureau — Census: July 2025 note with BTOS AI use updates
• Cost profile shifts, not new grants. With “no additional amounts authorized,” centers may have to reprioritize staff time toward AI curricula, potentially displacing other services or increasing reliance on third‑party trainers; GAO has previously flagged SBDC budgeting/administrative burdens when funding is tight or timing is uncertain. [9]U.S. Government Accountability Office — GAO-20-457: SBDCs—Reevaluation of how S…
• Risk of costly cyber incidents without basics. Rising breach costs (global average $4.88M in 2024) and DBIR findings on vulnerability exploitation and human error suggest SBDC curricula that couple AI tooling with NIST cybersecurity/AI‑risk practices could reduce downside. [10]IBM — IBM 2024 Cost of a Data Breach report press release[11]Verizon — 2024 Data Breach Investigations Report (DBIR) news release
• Compliance and marketing claims. FTC actions against deceptive AI claims (e.g., DoNotPay, writing‑tool cases) show legal exposure if firms overstate AI capabilities; SBDC guidance that stresses substantiation and fair‑marketing could avert losses. [12]Reuters — FTC announces crackdown on deceptive AI claims and schemes (coverage)[13]The Verge — FTC fines DoNotPay, targets deceptive AI claims (coverage)
• IP management. USPTO guidance (Feb 2024) clarifies that AI‑assisted inventions can be patentable where a human’s contribution is significant—relevant for SMBs integrating AI into R&D; SBDCs already advise on IP under 15 U.S.C. 648. [14]USPTO — USPTO issues inventorship guidance for AI‑assisted inventions (Feb. 12,…[3]Legal Information Institute (Cornell) — 15 U.S. Code § 648 - Small business dev…

Social Effects

• Worker impacts. Evidence shows AI tools can lift novice/low‑skill worker output disproportionately, potentially narrowing performance gaps within firms; however, tasks outside the studied domains may see smaller or negative effects. Training should emphasize augmentation, QA, and change management to avoid displacement. [6]National Bureau of Economic Research — Generative AI at Work (NBER Working Pape…
• Equity and access. BTOS trends indicate smaller firms adopt later; rural or under‑resourced areas may face broadband and skills gaps. Targeted SBDC outreach, remote modules, and toolkits can mitigate—but require capacity. [15]Web search · turn 2 #3
• Trust and consumer protection. AI‑enabled small businesses must handle personal data responsibly; FTC small‑business guidance (data minimization, vendor security, MFA, breach response) provides baselines that SBDCs can embed in trainings. [16]Web search · turn 5 #1[17]Web search · turn 5 #3[18]Web search · turn 5 #4
• Civil rights and responsible use. Aligning advice to the voluntary NIST AI RMF (and its GenAI profile) helps firms address bias, transparency, and safety without imposing new mandates. [4]NIST — NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0)[19]NIST — NIST AI RMF: Generative AI Profile (NIST-AI-600-1)

Environmental Effects

• Indirect energy demand. AI uptake by many small firms primarily relies on cloud services; at system level, IEA projects global data‑center electricity demand roughly doubling to ~945 TWh by 2030, with the US driving a large share—raising localized grid impacts even if any single SMB’s footprint is small. [20]International Energy Agency — IEA Energy and AI report – Executive summary[21]International Energy Agency — IEA Energy and AI – Energy demand from AI
• Local externalities. Data‑center growth concentrates in specific US clusters, where electricity and water use can strain local resources; firms should consider vendor disclosures on energy sourcing and water‑use efficiency when selecting AI/cloud tools. [20]International Energy Agency — IEA Energy and AI report – Executive summary
• Water footprint. Research highlights significant, often under‑reported water consumption tied to AI training/inference; while some providers report intensity reductions or zero‑water cooling at certain sites, these are provider claims and may vary by location. [22]arXiv — Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Fo…[23]Web search · turn 7 #0

Temporal Analysis

1. 0–12 months: Low‑cost interventions dominate—basic workshops on use‑case selection, prompt hygiene, data‑handling, vendor risk, and truthful claims; early productivity gains likely in customer support, marketing copy, and internal knowledge tasks. [6]National Bureau of Economic Research — Generative AI at Work (NBER Working Pape…[7]Stanford University — Experimental evidence on the productivity effects of gene…
2. 1–3 years: Divergence by sector and firm size; more workflows automated; heavier reliance on cloud AI heightens cumulative exposure to cyber, vendor lock‑in, and environmental externalities; adoption rates rise gradually from a low base among very small firms. [5]U.S. Census Bureau — Tracking Firm Use of AI in Real Time: A Snapshot from the…[20]International Energy Agency — IEA Energy and AI report – Executive summary
3. 3–5 years: Benefits depend on sustained skills development and controls aligned to NIST AI RMF/GenAI profile; firms that institutionalize risk management capture durable gains, laggards risk compliance failures or unproductive spend. [4]NIST — NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0)[19]NIST — NIST AI RMF: Generative AI Profile (NIST-AI-600-1)

Unintended Consequences

• Vendor capture/conflicts. AI training delivered with or by vendors risks implicit endorsements; SBDC networks must enforce conflict‑of‑interest policies under 13 CFR 130.320(g). [24]Web search · turn 14 #8
• Misleading AI marketing. SMBs adopting AI may overclaim capabilities; recent FTC actions show real enforcement risk—teach substantiation, avoid “robot lawyer”‑style claims. [12]Reuters — FTC announces crackdown on deceptive AI claims and schemes (coverage)[13]The Verge — FTC fines DoNotPay, targets deceptive AI claims (coverage)
• Shadow AI/data leakage. Unmanaged AI tools and plugins increase breach likelihood and cost; emphasize governance, access controls, and vendor due diligence consistent with FTC small‑business security guidance. [17]Web search · turn 5 #3[18]Web search · turn 5 #4
• Local environmental pushback. In data‑center‑dense regions, community concerns over energy/water use can spur permitting hurdles or higher utility costs that flow through to cloud pricing. SBDCs should brief clients on location‑specific risks when feasible. [20]International Energy Agency — IEA Energy and AI report – Executive summary
• Budget rules context. CUTGO limits apply to mandatory spending increases; this bill’s “no additional amounts authorized” language aligns with avoiding new mandatory costs but does not by itself ensure resource sufficiency at centers. [25]Congressional Research Service / Library of Congress — CRS: House Rule XXI, Cla…

Assessment

Overall stance: neutral. The proposal leverages an existing, nationwide advisory infrastructure and draws on an established federal AI definition. Evidence suggests real task‑level productivity upside, especially for smaller or less‑experienced teams, but realization depends on execution quality: vendor‑neutral curricula, integration of NIST AI‑risk and basic cybersecurity, and attention to uneven adoption and local environmental constraints. Absent new resources, scale and consistency of delivery remain the main risks. [2]U.S. Small Business Administration — Office of Small Business Development Cente…[1]Legal Information Institute (Cornell) — 15 U.S. Code § 9401 - Definitions (Nati…[6]National Bureau of Economic Research — Generative AI at Work (NBER Working Pape…[4]NIST — NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0)

Sourcing (key references)

Representative, verifiable sources underpinning the analysis.

• Statutory/program context: SBDC statute (15 U.S.C. 648) and AI definition (15 U.S.C. 9401). [3]Legal Information Institute (Cornell) — 15 U.S. Code § 648 - Small business dev…[1]Legal Information Institute (Cornell) — 15 U.S. Code § 9401 - Definitions (Nati…
• SBDC network scale and mission: SBA Office of Small Business Development Centers. [2]U.S. Small Business Administration — Office of Small Business Development Cente…
• Adoption data: Census BTOS working paper and updates (2023–2025). [5]U.S. Census Bureau — Tracking Firm Use of AI in Real Time: A Snapshot from the…[8]U.S. Census Bureau — Census: July 2025 note with BTOS AI use updates
• Productivity studies: NBER field study; Science‑published RCT summary. [6]National Bureau of Economic Research — Generative AI at Work (NBER Working Pape…[7]Stanford University — Experimental evidence on the productivity effects of gene…
• Risk frameworks: NIST AI Risk Management Framework and Generative AI Profile. [4]NIST — NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0)[19]NIST — NIST AI RMF: Generative AI Profile (NIST-AI-600-1)
• Cyber risk baseline: IBM Cost of a Data Breach 2024; Verizon DBIR 2024. [10]IBM — IBM 2024 Cost of a Data Breach report press release[11]Verizon — 2024 Data Breach Investigations Report (DBIR) news release
• Environmental context: IEA Energy and AI analysis (exec summary, demand module); AI water‑footprint research. [20]International Energy Agency — IEA Energy and AI report – Executive summary[21]International Energy Agency — IEA Energy and AI – Energy demand from AI[22]arXiv — Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Fo…
• Budget rules: CRS on the House CUTGO rule. [25]Congressional Research Service / Library of Congress — CRS: House Rule XXI, Cla…
• Program management risks: GAO on SBDC funding estimate burdens. [9]U.S. Government Accountability Office — GAO-20-457: SBDCs—Reevaluation of how S…