MétaCan
Menu
Back to cohort
Record W4412984931 · doi:10.1109/access.2025.3596060

AI Driven Fraud Detection Models in Financial Networks: A Comprehensive Systematic Review

2025· article· en· W4412984931 on OpenAlex

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

fundA Canadian funder is recorded on the work.
no affNo Canadian affiliation: this work is invisible to an affiliation-only frame.
No Canadian affiliation. An affiliation-only frame, the usual design, would never have seen this work. It is one of the works that make the case for inverting the frame.

Bibliographic record

VenueIEEE Access · 2025
Typearticle
Languageen
FieldComputer Science
TopicImbalanced Data Classification Techniques
Canadian institutionsnot available
FundersCanadian Institute for Advanced Research
KeywordsComputer scienceComputer security

Abstract

fetched live from OpenAlex

Rapid advancements in digital innovation and globalization has significantly increased the complexity of financial networks, making them more vulnerable to fraud. Traditional fraud detection methods struggle to keep pace with evolving fraudulent strategies, contributing to an estimated global financial loss of $ 5 trillion. In response, this review paper explores the role of artificial intelligence (AI) in financial fraud detection, highlighting machine learning (ML), deep learning (DL), and hybrid models as transformative solutions. By analyzing vast datasets, AI can uncover hidden fraud patterns and dynamically adapt to emerging threats. Techniques such as supervised and unsupervised learning, along with advanced approaches like Graph Neural Networks (GNNs), have proven particularly effective in detecting various types of financial fraud, including payment fraud, identity theft, and money laundering. This paper presents a comprehensive taxonomy of AI-driven fraud detection methodologies, synthesizing insights from a substantial number of research papers. It systematically categorizes fraud detection techniques based on their application in different types of fraud, providing a structured framework to understand their effectiveness. In addition, it examines the role of cloud computing, edge AI, and distributed systems in enabling real-time transaction monitoring and fraud detection. Although AI significantly improves detection accuracy, reduces operational costs, and strengthens regulatory compliance, challenges such as model explainability, data privacy concerns, algorithmic bias, and the dynamic nature of fraud remain critical barriers to widespread adoption. Our review highlights the need for collaborative efforts among financial institutions, regulators, and technology providers to address these challenges. Future research should focus on improving the transparency of the AI model, integrating AI with blockchain for secure data sharing, and leveraging federated learning to enhance fraud detection capabilities. By addressing these challenges, AI can play a pivotal role in securing financial systems, minimizing fraud risks, and fostering cross-industry collaboration for more resilient fraud detection frameworks.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.000
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Methods · Consensus signal: none
Teacher disagreement score0.989
Threshold uncertainty score0.558

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0000.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.001
Science and technology studies0.0000.000
Scholarly communication0.0000.002
Open science0.0020.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.028
GPT teacher head0.322
Teacher spread0.294 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it