FinTech’s gains and risks for commercial banks depend on who uses it and how: AI, blockchain and RWA enhance efficiency and innovation for digitally mature, large banks but can fuel superficial adoption and concentrated risk for smaller banks; regulators must adapt frameworks to manage technology-specific risk transmission.

Main Finding

The paper documents dual heterogeneity in how FinTech empowers commercial banks: (1) heterogeneity across technologies (AI, blockchain, RWA) in their channels, benefits and risks; and (2) heterogeneity across institutional endowments (bank size, digital foundation, resources, region) that conditions the magnitude and direction of effects. AI, blockchain and RWA each improve operational efficiency, risk management and product innovation in distinct ways, but also introduce technology-specific risk exposures (algorithmic discrimination for AI; decentralization–compliance tension for blockchain; valuation and risk‑stacking for RWA). Large banks tend to realize multi-technology collaborative innovation; small and medium banks more often adopt superficial or partial applications because of resource constraints. The paper argues for targeted FinTech resource allocation and adaptive regulatory/technical remedies (e.g., RegTech) to achieve coordinated development under controllable risk.

Key Points

• Conceptual framing

  • FinTech is defined as emergent business models and technology-enabled financial services (big data, cloud, AI, blockchain) that reshape pricing, delivery and risk assessment in banking.
  • The paper focuses on three concrete technologies: artificial intelligence (AI), blockchain, and real-world asset (RWA) digitization/valuation.

• Technology-specific effects

  • AI
    • Efficiency: improves credit scoring, risk prediction, customer segmentation, and personalized services; reduces costs and expands coverage.
    • Risks: algorithmic black boxes, data bias and algorithmic discrimination; opacity can amplify historical inequalities and create unfair resource allocation.
  • Blockchain
    • Efficiency: enables traceability, non‑tamperable transaction records, streamlined supply‑chain and receivables financing, and extended coverage to long‑tail SMEs.
    • Risks: decentralization vs. existing centralized supervisory frameworks; encryption/anonymity and cross‑jurisdictional ledgers complicate compliance and on‑chain supervision.
  • RWA (real‑world asset tokenization / valuation)
    • Efficiency: when RWA measurement/refined RWA regulation reduces risk weights and better aligns capital allocation, it can stimulate credit supply (esp. to SMEs) and improve capital use.
    • Risks: heterogeneous valuation capabilities across banks can create risk concentration or risk‑superposition (the paper notes valuation and transmission issues; discussion is partial).

• Heterogeneity by bank characteristics

  • Larger banks with richer data, R&D and capital can integrate multiple technologies and scale productive impact.
  • Smaller banks often lack digital foundations and resources; they adopt technologies superficially, limiting efficiency gains and possibly increasing operational or compliance risks.

• Policy and supervisory implications emphasized by the paper

  • Need for calibrated regulation that recognizes technology heterogeneity and bank heterogeneity.
  • Use of regulatory technology (RegTech) and supervisory adaptation to reconcile blockchain decentralization with territorial/real‑name oversight.
  • Guidance on FinTech resource allocation inside banks to avoid superficial deployments and manage risk exposures.

Data & Methods

• Primary approach: integrative literature review and logical deduction rather than original micro‑empirical estimation.
• Bibliometric analysis (papers 2010–2025)
  • Methods reported: keyword co‑occurrence mapping, cluster analysis, and “surprise detection” to identify evolving themes and research hotspots.
  • Findings from bibliometrics: evolution from cloud computing/big data (2010–2018) to digital finance, metaverse and business innovation (2020–2025); policy and practice coordination themes (inclusive finance, SME support, rural revitalization, financial risk).
• Comparative/qualitative synthesis
  • The paper synthesizes empirical findings from prior studies on profitability, risk transmission, product innovation and technology adoption.
  • Emphasizes mechanism-level analysis (how each technology affects efficiency, risk, and innovation) and conditional effects depending on bank size and digital maturity.
• Limitations (implicit)
  • No original causal identification or new bank-level panel regressions presented in the excerpts—analysis is conceptual and literature‑based.
  • RWA discussion in the provided text is incomplete (fragment ends mid‑argument), so some RWA mechanisms are summarized from partial content.

Implications for AI Economics

• For empirical research
  • Importance of heterogeneity: studies should explicitly model treatment effect heterogeneity by bank size, digital maturity, organizational resources and regional regulatory environments.
  • Recommended identification strategies:
    • Bank‑level panel differences‑in‑differences (policy or rollout shocks to FinTech adoption).
    • Instrumental variables for endogenous adoption (e.g., regional digital infrastructure rollout, incumbent tech partnerships).
    • Matching or synthetic controls for case studies of large versus small bank digital transformations.
    • Microdata needs: bank balance sheets, loan-level outcomes, internal risk scores, product usage logs, and measures of technical capability (IT spend, personnel).
  • Metrics to collect and report: productivity (cost-to-income, ROA/ROE), loan approval rates, time-to-service, non-performing loan (NPL) dynamics, credit access for SMEs, algorithmic fairness metrics (disparate impact), on‑chain activity measures.
• For theoretical and structural modeling
  • Incorporate tech‑specific production functions where AI reduces information frictions (improves match quality) but can introduce bias externalities; blockchain reduces verification costs but creates supervision frictions; RWA changes collateral valuation and capital constraints.
  • Model interactions: complementarities between technologies (AI + blockchain + RWA tokenization) and how they amplify or dampen each other’s benefits/risks.
  • Welfare trade-offs: quantify social gains from expanded credit and efficiency vs. distributional harms from algorithmic bias and systemic risk from opaque valuations.
• For policy and regulation
  • Design targeted, proportionate regulation recognizing:
    • Technology-specific risks: mandate model explainability and fairness audits for AI; on‑chain traceability standards and cross‑border data‑sharing protocols for blockchain; standardized valuation methodologies and disclosure for RWA.
    • Institution-specific constraints: support capacity building for SMEs and smaller banks (grants, shared infrastructure, sandboxed RegTech) to avoid a two‑tier system.
  • Promote RegTech adoption: invest in supervisory tools that can parse encrypted ledgers, audit smart contracts, and perform real‑time algorithmic monitoring.
• For practitioners and bank strategy
  • Prioritize foundational investments (data governance, talent, IT architecture) to realize sustained efficiency gains rather than piecemeal deployments.
  • Adopt multi‑technology pilots with clear metrics and phased scaling, combined with fairness and compliance checks.
• Open research directions highlighted
  • Causal evidence on how FinTech adoption alters credit allocation across firms and regions.
  • Measurement and mitigation of algorithmic discrimination in banking contexts.
  • Models and empirical work on how RWA tokenization affects systemic risk and capital adequacy.
  • Cross‑disciplinary studies linking metaverse/digital platforms, FinTech products, and inclusive finance outcomes.

Limitations of the paper: primarily a literature synthesis and conceptual analysis rather than new causal microeconometric evidence; parts of the RWA discussion are incomplete in the provided excerpt.