Authors

  • Sitora Toshpulatova
  • Nilufar Rashidova
  • G`ozal Ubaydullayeva
  • G Mr. Subhadhanuraja

DOI:

https://doi.org/10.71337/inlibrary.uz.science-research.86830

Keywords:

Fintech ecosystems artificial intelligence IMF World Bank BIS.

Abstract

This article explores how artificial intelligence (AI) is revolutionizing credit risk assessment in financial technology (Fintech) lending platforms. By analyzing large datasets in real-time, AI systems enhance accuracy, reduce bias, and expand access to credit. The research incorporates global insights from organizations like the IMF, World Bank, and BIS, along with examples from emerging Fintech ecosystems, including Uzbekistan. The paper also examines regulatory challenges, data privacy concerns, and practical recommendations for implementing AI in a responsible and inclusive way.

background image

ISSN:

2181-3906

2025

International scientific journal

«MODERN

SCIENCE

АND RESEARCH»

VOLUME 4 / ISSUE 5 / UIF:8.2 / MODERNSCIENCE.UZ

632

AI-POWERED RISK ASSESSMENT IN FINTECH LENDING PLATFORMS

Toshpulatova Sitora

1

BBA Student, Sambhram University, Jizzax, Uzbekistan.

Tel: +998 915995775

Rashidova Nilufar

BBA Student, Sambhram University, Jizzax, Uzbekistan.

Tel: +998 997696460

Ubaydullayeva G`ozal

BBA Student, Sambhram University, Jizzax, Uzbekistan.

Reg:21BA1001

Mr. Subhadhanuraja G

2

Assistant Professor, Department of Business Administration, Sambhram University.

Jizzax, Uzbekistan.

Email ID:

Subhadhanu625@gmail.com

https://doi.org/10.5281/zenodo.15389644

Abstract. This article explores how artificial intelligence (AI) is revolutionizing credit

risk assessment in financial technology (Fintech) lending platforms. By analyzing large datasets
in real-time, AI systems enhance accuracy, reduce bias, and expand access to credit. The
research incorporates global insights from organizations like the IMF, World Bank, and BIS,
along with examples from emerging Fintech ecosystems, including Uzbekistan. The paper also
examines regulatory challenges, data privacy concerns, and practical recommendations for
implementing AI in a responsible and inclusive way.

Key words: Fintech ecosystems, artificial intelligence, IMF, World Bank, BIS.

Introduction

The growth of Fintech lending platforms has reshaped the credit landscape by making

financing more accessible, especially to underbanked populations. Traditional credit scoring
models often rely on limited variables and historical data, which may exclude individuals
without formal financial histories.

Artificial Intelligence (AI) offers a new approach to risk assessment — one that is

dynamic, scalable, and based on real-time behavioral and alternative data. This paper examines
how AI is changing risk management practices in digital lending and what this means for the
future of financial inclusion and regulatory oversight.

2. Research Design
1. Type of Research

Qualitative and Exploratory:

The study focuses on identifying AI-driven innovations

and their influence on credit risk management.

Descriptive:

Aims to explain the technological transformation in risk evaluation rather

than measure impact quantitatively.

2. Data Collection

Secondary Data Sources:

Reports from the World Bank, IMF, BIS, academic journals,

Fintech whitepapers, and regulatory policy documents.


background image

ISSN:

2181-3906

2025

International scientific journal

«MODERN

SCIENCE

АND RESEARCH»

VOLUME 4 / ISSUE 5 / UIF:8.2 / MODERNSCIENCE.UZ

633

Case Studies:

Analysis of platforms such as Kabbage (USA), Tala (Kenya), and Uzum

Bank (Uzbekistan).

3. Method of Analysis

Thematic Analysis:

Focus on themes like algorithmic decision-making, credit scoring

alternatives, and bias mitigation.

Comparative Analysis:

Contrast between traditional financial institutions and AI-driven

Fintech models.

4. Scope and Limitations

Scope:

Focus on consumer and SME lending platforms employing AI for credit risk

assessment.

Limitations:

Lack of access to proprietary AI algorithms and limited empirical data on

post-lending performance in emerging markets.

Literature Review

AI-driven credit risk assessment has become a growing area of interest among scholars

and practitioners.

Traditional vs. AI Credit Models

Traditional models like FICO rely on limited variables. In contrast, AI models evaluate

hundreds of data points, including digital footprints, mobile usage, and social media behavior
(Jagtiani & Lemieux, 2018).

Inclusion through Alternative Data

World Bank (2022) reports that AI helps extend credit to populations previously excluded

by conventional scoring. This includes gig workers and informal sector participants.

Regulatory Perspectives

The Bank for International Settlements (2021) warns that while AI improves efficiency, it

also raises transparency and accountability concerns, especially when models are opaque ("black
box").

Bias and Fairness in AI Systems

O’Neil (2016) in

Weapons of Math Destruction

highlights how biased training data can

perpetuate inequality, stressing the need for ethical AI design.

Adoption in Emerging Markets

A report by the IMF (2023) shows Fintech adoption in Uzbekistan and Central Asia is

accelerating, with governments supporting AI experimentation in credit services.

The Role of AI in Risk Assessment
Enhanced Data Utilization

AI systems use structured and unstructured data — from transaction records to voice

analytics — to evaluate borrower profiles with higher granularity.

Real-Time Decision Making

Machine learning (ML) models enable instant loan approval based on live risk

assessments. This significantly reduces underwriting time and operational costs.

Fraud Detection and Anomaly Analysis

AI algorithms detect unusual patterns or fraud attempts faster than traditional systems.
This strengthens lender confidence and system integrity.


background image

ISSN:

2181-3906

2025

International scientific journal

«MODERN

SCIENCE

АND RESEARCH»

VOLUME 4 / ISSUE 5 / UIF:8.2 / MODERNSCIENCE.UZ

634

5. Risks and Challenges
5.1. Algorithmic Bias

Without careful design, AI systems may reflect historical inequalities in data. For

instance, if historical data favored urban borrowers, rural applicants may still be excluded.

5.2. Transparency and Explainability

Many AI models lack interpretability, making it difficult for users and regulators to

understand why a loan was denied.

5.3. Data Privacy

Collecting and processing vast personal data raises concerns about user privacy,

especially under GDPR and similar regulations.

5.4. Regulatory Gaps

Many jurisdictions lack clear guidelines on how AI-based credit decisions should be

regulated. This creates legal uncertainty for Fintech providers.

6. Case Examples
6.1. Tala (Kenya)

Uses AI to analyze SMS history, call records, and mobile payment behavior to assess

creditworthiness for unbanked users.

6.2. Kabbage (USA)

Employs real-time cash flow analysis from connected business accounts for SME loan

underwriting.

6.3. Uzum Bank (Uzbekistan)

Explores AI-driven credit scoring using utility bills, mobile behavior, and government e-

services integration.

7. Recommendations

Develop Ethical AI Standards:

Design AI systems with fairness and transparency in

mind.

Invest in Data Infrastructure:

Secure and interoperable data ecosystems are essential

for accurate assessments.

Foster Regulatory Sandboxes:

Allow experimentation under oversight to balance

innovation and protection.

Promote Financial and Digital Literacy:

Educate users on how AI decisions work and

their rights regarding data usage.

International Collaboration:

Share best practices across jurisdictions for harmonized

Fintech governance.

8. Conclusion

AI-powered credit risk assessment is transforming how loans are evaluated and

distributed in the Fintech ecosystem. While it improves access, speed, and efficiency, it also
raises ethical and regulatory challenges. With the right policies and technological safeguards, AI
can enable a more inclusive and resilient financial future.



background image

ISSN:

2181-3906

2025

International scientific journal

«MODERN

SCIENCE

АND RESEARCH»

VOLUME 4 / ISSUE 5 / UIF:8.2 / MODERNSCIENCE.UZ

635

REFERENCES

1.

Jagtiani, J., & Lemieux, C. (2018). The Roles of Alternative Data and Machine Learning
in Fintech Lending.

Federal Reserve Bank of Philadelphia

.

2.

O’Neil, C. (2016).

Weapons of Math Destruction

. Crown Publishing.

3.

World Bank (2022).

Digital Financial Inclusion Report

.

4.

Bank for International Settlements (2021).

Big tech in finance: opportunities and risks

.

5.

IMF (2023).

Fintech in Central Asia: Accelerating Access and Regulation

.

6.

Financial Stability Board (2022).

AI in Financial Services: Risks and Opportunities

.

7.

McKinsey & Company (2021).

Smart Credit Decisions with AI

.

8.

BIS (2023).

Supervising AI Models in Financial Institutions

.

9.

Uzum Bank Whitepaper (2024).

AI for Emerging Market Credit Models

.

10.

World Economic Forum (2020).

Shaping the Future of Financial Services

.

References

Jagtiani, J., & Lemieux, C. (2018). The Roles of Alternative Data and Machine Learning in Fintech Lending. Federal Reserve Bank of Philadelphia.

O’Neil, C. (2016). Weapons of Math Destruction. Crown Publishing.

World Bank (2022). Digital Financial Inclusion Report.

Bank for International Settlements (2021). Big tech in finance: opportunities and risks.

IMF (2023). Fintech in Central Asia: Accelerating Access and Regulation.

Financial Stability Board (2022). AI in Financial Services: Risks and Opportunities.

McKinsey & Company (2021). Smart Credit Decisions with AI.

BIS (2023). Supervising AI Models in Financial Institutions.

Uzum Bank Whitepaper (2024). AI for Emerging Market Credit Models.

World Economic Forum (2020). Shaping the Future of Financial Services.