The Commonplace
Home Papers Evidence Explore Trends Syntheses Digests References Docs 🎲 Workforce Futures
← Papers
Direction, evidence grade, and study type are AI-generated labels (gpt-5-mini), not human-verified. Syntheses are LLM-written. "Tensions" are machine-detected candidates, not confirmed contradictions. A research-acceleration tool, not peer review. How this is built →

AI-powered FinTech is linked to broader financial access in underserved U.S. neighborhoods, but benefits taper off after moderate penetration; community banks and credit unions that implement AI amplify inclusion more than purely digital or branch-only approaches.

Artificial Intelligence Fintech and the Interaction between Digital and Traditional Finance in Expanding Financial Inclusion among Underserved United States Communities
Iyedolapo Ajewole · August 08, 2026 · American Journal of Financial Technology and Innovation
openalex correlational low evidence 7/10 relevance Full text usable extracted full text DOI Source PDF

Structured author observations

Linked only from stored provider relations; the raw author line above is never matched by name.

OpenAlex

Latest observation:

  1. Iyedolapo Ajewole provider ID

Semantic Scholar

Latest observation:

  1. Iyedolapo Ajewole provider ID
Across 50,400 U.S. census tracts, greater AI-FinTech penetration is associated with higher financial inclusion, with diminishing marginal returns beyond an adoption index value of 0.387, and community banks/credit unions that adopt AI-enabled services produce multiplicative inclusion gains compared with digital-only or branch-only models.

Citation observations

Cumulative provider counts captured on specific dates; providers are never combined.

Financial exclusion continues to affect over 22 million unbanked and 47 million underbanked adults in underserving communities in the United States. The steep growth of artificial intelligence (AI)-enhanced financial technology (FinTech) platforms is an unprecedented opportunity to fill this gap by complementing, and not substituting, the traditional financial infrastructure. The paper will focus on the relationship between digital AI-FinTech solutions and conventional financial institutions in growing financial inclusion of underserved US communities, such as low-to-moderate income (LMI) households, rural communities, racial and ethnic minorities, and other historically marginalized communities. Based on the panel data of the Federal Deposit Insurance Corporation (FDIC) and the Federal Reserve Board, Consumer Financial Protection Bureau (CFPB) and Federal Communications Commission (FCC) surveys on 50,400 census tracts across three rounds of surveys (2015, 2019, 2023), this paper creates a composite Financial Inclusion Index based on a three-stage principal component analysis (PCA Findings affirm that AI-FinTech penetration has a strong and positive impact on financial inclusion; more so, there is a threshold effect, which is achieved at an index value of 0.387, beyond which the marginal returns to further AI-FinTech adoption become less significant and has policy implications on investment priority. The paper also shows that digital and traditional finance are complements: community banks and credit unions that implement AI-enabled services have a multiplicative inclusion effect relative to models based on digital-only and branches only. The policy proposals focus on the expansion of regulatory sandboxes, investment in broadband infrastructure, modernization of the Community Reinvestment Act (CRA), and algorithm fairness requirements.

Summary

Main Finding

AI-enhanced FinTech penetration significantly increases financial inclusion in underserved U.S. communities, but with a nonlinear (threshold) effect: once AI-FinTech penetration reaches an index value of 0.387 (AFTPI), marginal returns to further adoption decline. Moreover, AI-FinTech and traditional financial institutions (particularly community banks and credit unions that adopt AI) act as complements — their joint presence produces multiplicative inclusion gains relative to digital-only or branch-only models. Policy interventions (regulatory sandboxes, broadband investment, CRA modernization, algorithm-fairness rules) are necessary to realize and sustain these gains.

Key Points

  • Scope and problem: Financial exclusion remains concentrated — the paper cites over 22 million unbanked and 47 million underbanked adults — and is geographically and demographically clustered (LMI households, rural areas, racial/ethnic minorities, tribal lands).
  • Contributions:
  • Constructs a multidimensional Financial Inclusion Index (FII) at high geographic resolution (50,400 census tracts).
  • Documents a non-linear (threshold) relationship between AI-FinTech penetration and inclusion; identifies an AFTPI threshold = 0.387 beyond which marginal inclusion returns diminish.
  • Provides empirical evidence that AI-FinTech and traditional finance are complements; AI-enabled community banks and credit unions amplify inclusion more than digital-only providers or branches alone.
  • Translates findings into an action-oriented policy matrix across six domains (examples include sandboxes, broadband, CRA modernization, algorithm fairness).
  • Mechanisms: AI reduces unit costs of intermediation (alternative data, ML underwriting, real-time decisions), enabling cost-effective outreach to thin-file/no-file consumers. However, systemic constraints (digital divide, digital literacy, consumer trust, regulatory design) can be binding.
  • Structural caveats: The digital divide — lack of broadband and device access — limits AI-FinTech uptake in many underserved tracts; historical and structural discrimination in banking shapes baseline exclusion.

Data & Methods

  • Data:
    • Panel of U.S. census-tract-level data (50,400 tracts) combining FDIC, Federal Reserve Board, CFPB, and FCC survey sources.
    • Three survey rounds: 2015, 2019, 2023.
    • Supplementary sector-level adoption metrics (e.g., CCAF & WEF: AI adoption in U.S. FinTech rose from 41% in 2019 to 78% in 2023).
  • Outcome construction:
    • Financial Inclusion Index (FII): a composite, multidimensional index capturing access and use of both digital and traditional financial services, created via a three-stage principal component analysis (PCA).
  • Empirical approach (as reported):
    • Panel econometric analysis linking AI-FinTech penetration measures to the FII.
    • Nonlinear/threshold estimation to detect the AFTPI = 0.387 threshold where marginal returns to AI-FinTech adoption decline.
    • Tests of interaction/complementarity between AI-FinTech penetration and presence/adoption of AI-enabled community banks/credit unions (showing multiplicative effects vs. single-channel models).
  • Controls and concerns: Models incorporate contextual constraints (broadband access, digital literacy proxies, regulatory environment) and exploit fine geographic granularity to address heterogeneity; the paper situates results in systemic theory to highlight omitted-system risks.

Implications for AI Economics

  • Returns-to-AI are nonlinear and context-dependent: The identified threshold (AFTPI = 0.387) implies diminishing marginal social returns past a certain market penetration — informing optimal sequencing and targeting of public and private AI-FinTech investments.
  • Complementarities matter: Policies and business strategies that encourage partnerships between AI-FinTech firms and community financial institutions can generate larger inclusion and welfare gains than digital-only deployment. Economic models of AI adoption should incorporate interaction effects with incumbent institutions.
  • Distributional and welfare modeling: AI-FinTech can disproportionately benefit historically excluded groups, but only when complementary subsystems (broadband, digital literacy, trust-building, fair regulation) are addressed. Welfare analysis of AI deployment must include these binding constraints and uneven access to digital infrastructure.
  • Policy design and regulation:
    • Investment priorities: Given the threshold, prioritize expanding AI-FinTech in tracts below the AFTPI and invest in infrastructure (broadband, ACP) and capability (digital literacy) to raise effectiveness.
    • Regulatory tools: Expand regulatory sandboxes to test inclusive AI products, modernize CRA to account for digital outreach, and implement algorithm-fairness/monitoring requirements to prevent discriminatory outcomes from alternative-data models.
  • Research directions for AI economics:
    • Quantify dynamic welfare gains from complementary deployments (AI-FinTech + community banks) versus substitutes.
    • Model equilibrium provision of digital infrastructure and financial services to capture feedbacks between broadband adoption and FinTech uptake.
    • Study long-run distributional impacts and intergenerational wealth effects of AI-enabled credit access in historically excluded communities.
    • Evaluate cost-benefit thresholds of public subsidies (broadband, ACP) that unlock higher marginal returns to AI-FinTech investment.

If you want, I can (a) extract and summarize the exact variables included in the FII and the AI-FinTech penetration measure from the paper, (b) outline the likely econometric specification(s) used (fixed effects, threshold regression, interaction terms) in greater technical detail, or (c) produce a short policy brief targeted to regulators or community banks. Which would you prefer?

Assessment

Paper Typecorrelational Evidence Strengthlow — Large, fine-grained panel and a composite outcome are strengths, but causal claims rely on observational associations without a clearly stated exogenous source of variation or formal identification strategy to rule out confounding, reverse causality, or selection into AI-FinTech adoption. Methods Rigormedium — The paper uses a large census-tract panel (50,400 tracts), a multi-stage PCA to build a multidimensional Financial Inclusion Index, and explores non-linearity/thresholds and complementarities with traditional banks—these are appropriate and informative choices. However, the supplied text does not document strong strategies for addressing endogeneity (e.g., instruments, difference-in-differences with exogenous shocks), measurement validation of AI-FinTech penetration, or robustness checks in detail, which limits causal interpretation. SamplePanel of 50,400 U.S. census tracts using three rounds of national surveys/data sources (FDIC, Federal Reserve Board, CFPB, FCC) from 2015, 2019, and 2023; constructs tract-level Financial Inclusion Index via three-stage PCA and uses measures of AI-FinTech penetration, broadband access, demographic and socioeconomic covariates, and institutional variables (e.g., local community bank/credit union activity). Themesadoption inequality IdentificationObservational panel analysis using cross-census-tract variation over three survey rounds (2015, 2019, 2023); constructs a composite Financial Inclusion Index via three-stage PCA and relates it to a measure of AI-FinTech penetration using multivariate regression including controls and non-linear (threshold/segmented) specifications to detect diminishing marginal returns; no clearly described quasi-experimental source (instrument, policy discontinuity, or randomized variation) is reported in the supplied text. GeneralizabilityResults are US-specific and may not generalize to developing countries or different regulatory contexts., Census-tract aggregates may mask within-tract heterogeneity and individual-level behavioral differences., Measures of AI-FinTech penetration and the composite Financial Inclusion Index may be sensitive to indicator selection and PCA weighting., Findings pertain to 2015–2023 survey rounds and may not capture rapid post-2023 changes in AI deployment or policy., Observational design limits external validity for causal policy inference (e.g., scaling strategies may perform differently under alternative market or regulatory conditions).

Claims (9)

ClaimDirectionOutcomeConfidence & EvidenceDetails
AI-FinTech penetration has a strong and positive impact on financial inclusion in underserved United States communities. Consumer Welfare positive Composite Financial Inclusion Index
Reading fidelity high
Study strength medium
n=50400
0.3
The relationship between AI-FinTech penetration and financial inclusion exhibits a threshold effect: after the AI-FinTech penetration index reaches 0.387, the marginal returns to additional adoption become less significant. Consumer Welfare positive Marginal effect of AI-FinTech penetration on financial inclusion
Reading fidelity high
Study strength medium
n=50400
index value of 0.387
0.3
Digital finance and traditional finance are complements rather than substitutes in promoting financial inclusion. Consumer Welfare positive Financial inclusion associated with the interaction of digital and traditional finance
Reading fidelity high
Study strength medium
n=50400
0.3
Community banks and credit unions that implement AI-enabled services produce a multiplicative financial-inclusion effect relative to digital-only and branch-only models. Consumer Welfare positive Financial inclusion effect of combined AI-enabled and traditional financial services
Reading fidelity high
Study strength medium
n=50400
0.3
The United States has 5.9 million completely unbanked households and 18.7 million underbanked households that rely mainly on high-cost alternative financial services. Consumer Welfare negative Number of unbanked and underbanked households
Reading fidelity high
Study strength high
5.9 million unbanked households; 18.7 million underbanked households
0.5
Financial exclusion is substantially higher among Black and Hispanic households than among White non-Hispanic households. Inequality negative Share of households that are unbanked or underbanked by racial and ethnic group
Reading fidelity high
Study strength high
36.7% of Black households; 28.1% of Hispanic households; 7.2% of White non-Hispanic households
0.5
Households dependent on check cashers and payday lenders spend approximately $189 per year on transaction charges. Consumer Welfare negative Annual transaction costs paid by financially excluded households
Reading fidelity high
Study strength medium
$189 a year
0.3
Approximately 21.3 million Americans lack broadband access at speeds sufficient to use modern FinTech applications, with low-income urban neighborhoods, tribal areas, and rural areas disproportionately affected. Automation Exposure negative Access to broadband infrastructure suitable for FinTech use
Reading fidelity high
Study strength medium
21.3 million Americans
0.3
AI adoption among U.S. FinTech firms increased from 41% in 2019 to 78% in 2023, with the highest adoption levels in lending and insurance underwriting. Adoption Rate positive AI adoption rate among U.S. FinTech firms
Reading fidelity high
Study strength medium
78% adopting AI in 2023 compared to 41% in 2019
0.3

Notes