Journal of Economic, Finance Research and Review (JEFRR)

The Double Divide Revisited: Sector Composition, Within-Sector Adoption and Artificial Intelligence Use Among Minority- And Women-Owned U.S. Businesses

Oni, Olufemi Olumayowa, Usman, Abimbola, Oni, Olusegun Opeyemi, Okwarajiaku, Promise

28 October 2026 · Vol. 2, Issue 10, pp. 728-742

DOI: 10.65150/EP-jefrr/V2E10/2026-05

Abstract

Purpose. Policy discussion often assumes that minority- and women-owned businesses face a “double divide” in artificial intelligence (AI) adoption, compounding their other disadvantages. This study tests that assumption. It asks how far owner-group differences in AI use reflect the sectors in which firms operate (composition) and how far they reflect differences in adoption among firms in the same sector. Methodology. Drawing on mixed-embeddedness and resource-constraint theories of minority entrepreneurship, the study combines six published tabulations from the U.S. Census Bureau’s Annual Business Survey (ABS) for reference years 2017 and 2022 with sector adoption rates from the Business Trends and Outlook Survey (BTOS) AI Supplement for 2025–26. Owner-group gaps are decomposed by indirect standardisation (a Kitagawa-type decomposition) into a sector-composition component and a within-sector component, and a size-composition component is estimated separately. Monte Carlo confidence intervals are derived from the published standard errors. A forward-looking exposure index applies 2025–26 BTOS sector rates to the 2022 sector composition and compares the result with 2017 rates over the same ten sectors. Findings. Sector composition places Hispanic-, Asian- and minority-owned firms at a disadvantage: in 2017 their composition-implied AI use was 6–8% below that of the baseline group. Net of sector composition, however, minority-owned firms used AI more often than nonminority firms (+1.09 percentage points, 95% CI 0.81–1.38), and this advantage outweighed the composition penalty for every minority group. In 2022 minority-owned firms again reported higher use (3.8% vs. 3.1%, p = .002); the approximate 2022 within-sector component remained positive and significant for minority-, Hispanic- and Black-owned firms. Minority-owned firms were also 7.3 points less likely to report that AI was “not applicable” to their business. Women-owned firms show the opposite profile: a favourable sector mix but a within-sector shortfall in 2017 (−0.53 points) that was no longer statistically detectable in 2022. Compared over the same ten sectors, the relative composition penalty of Hispanic- and Asian-owned firms widens from about 9–10% under 2017 sector rates to about 12% under 2025–26 rates. Originality. This is the first decomposition of owner-group AI-adoption gaps using official U.S. statistics. It shows that, at the level of published aggregates, the “double divide” for minority-owned firms is a sector-placement divide rather than a shortfall in within-sector adoption, and that the placement disadvantage is larger under current sector adoption rates than under earlier ones.

Keywords: artificial intelligence, minority entrepreneurship, women-owned businesses, technology adoption, decomposition, Annual Business Survey

Read the full text on Journal of Economic, Finance Research and Review →The full peer-reviewed article and PDF are hosted on the journal's site (the version of record).

Oni, Olufemi Olumayowa, Usman, Abimbola, Oni, Olusegun Opeyemi, Okwarajiaku, & Promise (2026). The Double Divide Revisited: Sector Composition, Within-Sector Adoption and Artificial Intelligence Use Among Minority- And Women-Owned U.S. Businesses. Journal of Economic, Finance Research and Review, 2(10), 728-742. https://doi.org/10.65150/EP-jefrr/V2E10/2026-05

@article{Oni2026,
  title   = {The Double Divide Revisited: Sector Composition, Within-Sector Adoption and Artificial Intelligence Use Among Minority- And Women-Owned U.S. Businesses},
  author  = {Oni and Olufemi Olumayowa and Usman and Abimbola and Oni and Olusegun Opeyemi and Okwarajiaku and Promise},
  journal = {Journal of Economic, Finance Research and Review},
  year    = {2026},
  volume  = {2},
  number  = {10},
  pages   = {728-742},
  doi     = {10.65150/EP-jefrr/V2E10/2026-05},
  url     = {https://doi.org/10.65150/EP-jefrr/V2E10/2026-05}
}

← Back to all articles