119-HR-5784 Investigative Journalist Impact Analysis

Summary

H.R. 5784 (AI-WISE Act) directs the Small Business Administration (SBA) to publish public, vendor-neutral AI education modules for small businesses within 180 days, drawing on NIST and an SBA advisory group, with no new funds authorized. The bill’s scope is informational (not regulatory). Expected impacts: improved SME AI literacy and decision quality; small administrative costs absorbed by SBA; indirect environmental effects via downstream AI adoption; and governance risks from an advisory group exempt from the Federal Advisory Committee Act (FACA). [1]Congress.gov / Library of Congress — All Information (Except Text) for H.R.5784…

Implementation leans on the NIST AI Risk Management Framework and SBA’s existing Learning Platform and Resource Partner network (SBDCs, SCORE, WBCs, VBOCs). [2]NIST — Artificial Intelligence Risk Management Framework (AI RMF 1.0)[3]U.S. Small Business Administration — SBA Learning Platform[4]U.S. Small Business Administration — SBA Resource Partners

Economic Effects

Evidence-driven estimates of business, income, assets, employment, and markets.

• Administrative cost to SBA: Minimal incremental spend because the bill authorizes no additional amounts; costs are opportunity costs within SBA’s existing training budgets and staff. Absence of a CBO estimate as of Nov 20, 2025 prevents a quantified score. [1]Congress.gov / Library of Congress — All Information (Except Text) for H.R.5784…
• SME productivity potential: Controlled experiments show a 14% average productivity lift from generative AI assistance in customer support, with largest gains for less-experienced workers—consistent with the bill’s emphasis on “actionable” literacy. Education that helps firms identify tasks AI can reliably perform is likely to unlock such gains prudently. [5]NBER — Generative AI at Work (NBER Working Paper 31161)[6]Web search · turn 1 #1
• Adoption baseline: Census BTOS data indicate AI use among U.S. businesses was low but rising (e.g., ~3.8–5.4% in late-2023/early-2024). Education targeted at evaluation and risk may raise effective (not just nominal) adoption among SMEs. [7]U.S. Census Bureau — How Many U.S. Businesses Use Artificial Intelligence? (BTO…
• SME demand for guidance: OECD 2025 survey finds 31% of SMEs use generative AI, but many cite legal/privacy and capability concerns—areas the modules are designed to address—suggesting tangible demand for neutral, risk-aware training. [8]OECD — Generative AI and the SME Workforce: New Survey Evidence
• Market neutrality: Statutory language prohibits preference for particular AI tools or vendors; if enforced, this reduces risk of implicit endorsements and vendor lock-in, supporting fair competition across providers. [6]Web search · turn 1 #1

Social Effects

Implications for communities, demographic groups, and vulnerable populations.

• Workforce upskilling: By emphasizing “how models work,” limits, and human-in-the-loop decision-making, modules can improve front-line decision quality and reduce harm from overreliance—especially helpful for microbusinesses with thin technical capacity. [6]Web search · turn 1 #1
• Consumer protection spillovers: FTC has warned businesses and AI providers about privacy/compliance and deceptive claims. SBA modules that teach privacy-by-design and truthful marketing can reduce exposure for small firms and their customers. [9]Federal Trade Commission — AI Companies: Uphold Your Privacy and Confidentialit…
• Fraud resilience: SMEs are frequent targets of cyber-enabled fraud (e.g., business email compromise). Training on detecting AI-generated content and maintaining human checks may mitigate losses in a high-risk landscape. [10]FBI — FBI’s Internet Crime Complaint Center Annual Report Released for 2023
• Equity: Leveraging existing SBA Resource Partners (SBDCs, WBCs, VBOCs, SCORE) helps reach underserved groups; localized modules can tailor to community needs if neutrality safeguards are maintained. [4]U.S. Small Business Administration — SBA Resource Partners

Environmental Effects

Sustainability, resource use, emissions, and ecological impact.

• Indirect load growth: Education that accelerates AI uptake may contribute, at the margin, to rising data-center electricity demand. IEA projects data-center electricity use to roughly double by 2030, with AI-optimized servers a major driver and pronounced local grid impacts in the U.S. [11]International Energy Agency — Energy demand from AI – Energy and AI
• Efficiency upside: The same literacy may steer SMEs toward lower-footprint use cases (e.g., building energy management, logistics optimization) that studies associate with meaningful energy savings, partially offsetting upstream compute loads. [12]Web search · turn 11 #0
• Net environmental balance: Because this bill funds information rather than infrastructure, effects depend on module content and adoption patterns; pairing guidance on carbon/energy accounting with case studies of efficiency-focused AI could nudge reductions. [6]Web search · turn 1 #1

Temporal Analysis

Short-term vs. long-term outcomes.

1. 0–6 months post-enactment: SBA stands up public modules on its learning platform; early benefits are awareness-raising and basic risk hygiene (privacy, data handling, human oversight). Measurable firm-level productivity or risk reductions are limited at this stage. [6]Web search · turn 1 #1[3]U.S. Small Business Administration — SBA Learning Platform
2. 6–24 months: As Resource Partners localize delivery and SMEs apply guidance to evaluate pilots, expect selective productivity gains (especially in customer-facing and back-office tasks) and fewer costly missteps (e.g., mishandled data, deceptive claims). [4]U.S. Small Business Administration — SBA Resource Partners[5]NBER — Generative AI at Work (NBER Working Paper 31161)[9]Federal Trade Commission — AI Companies: Uphold Your Privacy and Confidentialit…
3. 2+ years: Effects hinge on module upkeep and neutrality enforcement. If the curriculum stays current with NIST RMF updates and legal developments (copyright, privacy), longer-run benefits compound; if not, guidance decays and risk rises. [2]NIST — Artificial Intelligence Risk Management Framework (AI RMF 1.0)[13]Web search · turn 10 #0

Unintended Consequences

Credible risks, trade-offs, or side effects.

• Transparency gap: The advisory working group is explicitly exempt from FACA. Without FACA-style disclosures, membership, deliberations, and conflicts may be opaque—raising vendor-capture risk in curriculum content. Mitigation: publish rosters, meeting notes, and COI statements voluntarily. [6]Web search · turn 1 #1
• Privacy/compliance pitfalls: If modules underplay data-handling risks, SMEs may inadvertently expose customer or proprietary data to model-as-a-service providers, violating promises and triggering FTC action. Training must emphasize contract terms, data retention, and off-switches. [9]Federal Trade Commission — AI Companies: Uphold Your Privacy and Confidentialit…
• False confidence/hallucinations: Literacy without guardrails could increase overreliance on AI outputs. NIST RMF emphasizes human oversight and continuous monitoring; modules should embed these practices. [2]NIST — Artificial Intelligence Risk Management Framework (AI RMF 1.0)
• Opportunity cost at SBA: “No additional amounts authorized” implies reallocation. If under-resourced, modules may be static or generic, eroding trust and usefulness. Progress should be benchmarked and refreshed against NIST updates and evolving law (e.g., copyright). [1]Congress.gov / Library of Congress — All Information (Except Text) for H.R.5784…[2]NIST — Artificial Intelligence Risk Management Framework (AI RMF 1.0)[13]Web search · turn 10 #0
• Cyber-enabled fraud amplification: As AI gets easier to use, attacker capability rises. Modules should cover BEC defenses, deepfake detection, and verification workflows; otherwise, net losses to SMEs may persist. [10]FBI — FBI’s Internet Crime Complaint Center Annual Report Released for 2023

Assessment

Analytical summary (not advocacy).

Overall stance: Neutral. The bill is a low-cost, education-first intervention that—if implemented with rigorous neutrality and updated against NIST RMF and evolving legal guidance—can modestly improve SME productivity and reduce avoidable AI-related risks. Environmental effects are indirect and hinge on use cases promoted. Key governance weaknesses (FACA exemption, funding constraints) warrant oversight to prevent capture and ensure the curriculum stays current and balanced. [1]Congress.gov / Library of Congress — All Information (Except Text) for H.R.5784…[2]NIST — Artificial Intelligence Risk Management Framework (AI RMF 1.0)

Sourcing and Key References

Selected authoritative materials used in this assessment.

• Bill text, titles, actions, and status for H.R. 5784 (AI-WISE Act), Congress.gov. [1]Congress.gov / Library of Congress — All Information (Except Text) for H.R.5784…[14]Web search · turn 1 #3
• NIST AI Risk Management Framework and resources (AI RMF 1.0; Playbook; AIRC). [2]NIST — Artificial Intelligence Risk Management Framework (AI RMF 1.0)
• SBA Learning Platform and Resource Partner network descriptions. [3]U.S. Small Business Administration — SBA Learning Platform[4]U.S. Small Business Administration — SBA Resource Partners
• U.S. Census Bureau BTOS: baseline AI adoption and methodology. [7]U.S. Census Bureau — How Many U.S. Businesses Use Artificial Intelligence? (BTO…
• OECD (2025) Generative AI and the SME Workforce: new survey evidence (use, barriers, perceived benefits). [8]OECD — Generative AI and the SME Workforce: New Survey Evidence
• NBER evidence on productivity effects from generative AI assistance. [5]NBER — Generative AI at Work (NBER Working Paper 31161)
• IEA Energy and AI: data center electricity demand projections and local impacts. [11]International Energy Agency — Energy demand from AI – Energy and AI
• FTC guidance on privacy/confidentiality and AI providers’ commitments. [9]Federal Trade Commission — AI Companies: Uphold Your Privacy and Confidentialit…
• FBI IC3 indicators on business email compromise and cyber-enabled fraud. [10]FBI — FBI’s Internet Crime Complaint Center Annual Report Released for 2023
• Definition of “artificial intelligence” in 15 U.S.C. § 9401 (incorporated by reference in the bill). [15]Legal Information Institute (Cornell) — 15 U.S. Code § 9401 - Definitions