Generative AI in finance: Replicability, methodological contingencies, and future research directions

Hassnian Ali; Muhammad Bilal Zafar; Ahmet Faruk Aysan

doi:10.1016/j.frl.2025.108797

Generative AI in finance: Replicability, methodological contingencies, and future research directions

Hassnian Ali^*
, Muhammad Bilal Zafar
, Ahmet Faruk Aysan

^*Corresponding author for this work

Islamic Studies – Administration

Research output: Contribution to journal › Article › peer-review

1 Citation (Scopus)

Abstract

Generative Artificial Intelligence (AI) is reshaping finance by transforming decision-making, risk management, and stakeholder engagement. This study provides a theory-informed synthesis of 84 peer-reviewed articles (2022-2025) using PRISMA-based screening, bibliometric analysis, and Structural Topic Modeling (STM). Six themes emerge: financial decision-making, ESG analytics, stock market prediction, advanced modeling for fraud detection and explainable AI, ChatGPT in accounting and education, and sentiment analysis with domain-specific LLMs. Findings show that generative AI enhances predictive capabilities and ESG assessments but raises issues of bias, transparency, and regulation. The review outlines future research priorities around interpretability, multimodal data, and governance frameworks.

Original language	English
Article number	108797
Journal	Finance Research Letters
Volume	86
DOIs	https://doi.org/10.1016/j.frl.2025.108797
Publication status	Published - 1 Dec 2025

Keywords

Bibliometric
Finance
Generative AI
Structural topic modeling
Systematic literature review

Access to Document

10.1016/j.frl.2025.108797

Cite this

@article{8d2ffc2fdd664857a797b2e9fe222bd6,

title = "Generative AI in finance: Replicability, methodological contingencies, and future research directions",

abstract = "Generative Artificial Intelligence (AI) is reshaping finance by transforming decision-making, risk management, and stakeholder engagement. This study provides a theory-informed synthesis of 84 peer-reviewed articles (2022-2025) using PRISMA-based screening, bibliometric analysis, and Structural Topic Modeling (STM). Six themes emerge: financial decision-making, ESG analytics, stock market prediction, advanced modeling for fraud detection and explainable AI, ChatGPT in accounting and education, and sentiment analysis with domain-specific LLMs. Findings show that generative AI enhances predictive capabilities and ESG assessments but raises issues of bias, transparency, and regulation. The review outlines future research priorities around interpretability, multimodal data, and governance frameworks.",

keywords = "Bibliometric, Finance, Generative AI, Structural topic modeling, Systematic literature review",

author = "Hassnian Ali and Zafar, \{Muhammad Bilal\} and Aysan, \{Ahmet Faruk\}",

note = "Publisher Copyright: {\textcopyright} 2025 The Author(s).",

year = "2025",

month = dec,

day = "1",

doi = "10.1016/j.frl.2025.108797",

language = "English",

volume = "86",

journal = "Finance Research Letters",

issn = "1544-6123",

publisher = "Elsevier Ltd",

}

TY - JOUR

T1 - Generative AI in finance

T2 - Replicability, methodological contingencies, and future research directions

AU - Ali, Hassnian

AU - Zafar, Muhammad Bilal

AU - Aysan, Ahmet Faruk

PY - 2025/12/1

Y1 - 2025/12/1

N2 - Generative Artificial Intelligence (AI) is reshaping finance by transforming decision-making, risk management, and stakeholder engagement. This study provides a theory-informed synthesis of 84 peer-reviewed articles (2022-2025) using PRISMA-based screening, bibliometric analysis, and Structural Topic Modeling (STM). Six themes emerge: financial decision-making, ESG analytics, stock market prediction, advanced modeling for fraud detection and explainable AI, ChatGPT in accounting and education, and sentiment analysis with domain-specific LLMs. Findings show that generative AI enhances predictive capabilities and ESG assessments but raises issues of bias, transparency, and regulation. The review outlines future research priorities around interpretability, multimodal data, and governance frameworks.

AB - Generative Artificial Intelligence (AI) is reshaping finance by transforming decision-making, risk management, and stakeholder engagement. This study provides a theory-informed synthesis of 84 peer-reviewed articles (2022-2025) using PRISMA-based screening, bibliometric analysis, and Structural Topic Modeling (STM). Six themes emerge: financial decision-making, ESG analytics, stock market prediction, advanced modeling for fraud detection and explainable AI, ChatGPT in accounting and education, and sentiment analysis with domain-specific LLMs. Findings show that generative AI enhances predictive capabilities and ESG assessments but raises issues of bias, transparency, and regulation. The review outlines future research priorities around interpretability, multimodal data, and governance frameworks.

KW - Bibliometric

KW - Finance

KW - Generative AI

KW - Structural topic modeling

KW - Systematic literature review

UR - https://www.scopus.com/pages/publications/105022836690

U2 - 10.1016/j.frl.2025.108797

DO - 10.1016/j.frl.2025.108797

M3 - Article

AN - SCOPUS:105022836690

SN - 1544-6123

VL - 86

JO - Finance Research Letters

JF - Finance Research Letters

M1 - 108797

ER -

Generative AI in finance: Replicability, methodological contingencies, and future research directions

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this