GAO finds SBA failed to consistently disclose AI use cases for years

The SBA failed to publicly report its AI use for six years, violating federal requirements in place since 2020, a GAO report found. The agency used AI for fraud detection and pandemic loan screening but never disclosed it.

Categorized in: AI News Government
Published on: May 06, 2026
GAO finds SBA failed to consistently disclose AI use cases for years

SBA Failed to Report AI Use for Six Years, GAO Finds

The Small Business Administration did not publicly disclose how it uses artificial intelligence until March 2026, despite federal requirements to do so since 2020, according to a Government Accountability Office report released Monday.

Federal agencies must report specific AI use cases-including how systems are designed, developed, and procured-to the public. SBA's six-year gap in compliance drew congressional scrutiny after Sen. Joni Ernst, R-Iowa, questioned why the agency had not reported AI applications it was openly using, such as fraud detection in loan programs and pandemic relief screening.

Why SBA Missed the Deadline

Between 2021 and 2023, SBA officials said the agency had no reportable AI use cases. In 2024, SBA created an internal inventory but never published it, citing inadequate documentation and staff turnover.

As of April 2026, SBA said its AI policies remain under revision and reporting processes are still being developed.

What GAO Wants SBA to Do

The GAO called on SBA to establish policies and procedures that ensure consistent compliance with public reporting requirements. The agency must also define roles and responsibilities for reporting and document how it implements those policies.

SBA agreed with the recommendations.

Where SBA Could Deploy AI

The GAO identified several potential uses for AI at SBA: conducting market research, reviewing proposals, analyzing submitted data, drafting reports, and detecting fraud.

The agency oversees small business programs across the federal government, making data management and reporting particularly labor-intensive. AI could reduce that workload.

But the GAO warned of real risks. AI systems can produce inaccurate outputs, create data privacy and security vulnerabilities, and generate biased results that affect decision-making. Agencies also face implementation challenges, including limited technical expertise and complex approval processes for new technologies.

Learn more about AI for Government and how federal agencies are approaching technology adoption.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)