Nos derniers articles

This paper 𝗲𝘅𝗮𝗺𝗶𝗻𝗲𝘀 𝘁𝗵𝗲 𝗲𝘀𝗰𝗮𝗹𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗿𝗲𝗮𝘁 𝗼𝗳 𝗔𝗜-𝗱𝗿𝗶𝘃𝗲𝗻 𝗳𝗿𝗮𝘂𝗱 𝗮𝗻𝗱 𝗰𝘆𝗯𝗲𝗿𝗰𝗿𝗶𝗺𝗲, highlighting how criminal organizations are rapidly adopting advanced AI, particularly generative AI, to execute sophisticated attacks. It details how these malicious uses lead to 𝗶𝗻𝗰𝗿𝗲𝗮𝘀𝗲𝗱 𝗳𝗶𝗻𝗮𝗻𝗰𝗶𝗮𝗹 𝗹𝗼𝘀𝘀𝗲𝘀, 𝗺𝗼𝗿𝗲 𝗶𝗻𝘁𝗿𝗶𝗰𝗮𝘁𝗲 𝗰𝗿𝗶𝗺𝗲 𝗽𝗮𝘁𝘁𝗲𝗿𝗻𝘀, 𝗮𝗻𝗱 𝗻𝗼𝘃𝗲𝗹 𝘀𝗰𝗮𝗺 𝘁𝘆𝗽𝗼𝗹𝗼𝗴𝗶𝗲𝘀, such as deepfakes and advanced phishing. The document also 𝗲𝘅𝗽𝗹𝗼𝗿𝗲𝘀 𝘁𝗵𝗲 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝘀 𝗳𝗮𝗰𝗲𝗱 𝗯𝘆 𝗳𝗶𝗻𝗮𝗻𝗰𝗶𝗮𝗹 𝗶𝗻𝘀𝘁𝗶𝘁𝘂𝘁𝗶𝗼𝗻𝘀 𝗶𝗻 𝗱𝗲𝗳𝗲𝗻𝗱𝗶𝗻𝗴 𝗮𝗴𝗮𝗶𝗻𝘀𝘁 𝘁𝗵𝗲𝘀𝗲 𝘁𝗵𝗿𝗲𝗮𝘁𝘀, citing issues like slow AI adoption, outdated risk management frameworks, and underinvestment in defense systems. Ultimately, it 𝗮𝗱𝘃𝗼𝗰𝗮𝘁𝗲𝘀 𝗳𝗼𝗿 𝘁𝗵𝗲 𝘂𝗿𝗴𝗲𝗻𝘁 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 𝗼𝗳 𝗮𝗴𝗶𝗹𝗲, 𝗔𝗜-𝘃𝗲𝗿𝘀𝘂𝘀-𝗔𝗜 𝗱𝗲𝗳𝗲𝗻𝘀𝗲 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗲𝘀 and emphasizes the critical need for industry-wide cooperation to counteract the evolving landscape of AI-enabled financial crime.

This opinion and accompanying report from the 𝗘𝗕𝗔 provides a comprehensive overview of 𝗺𝗼𝗻𝗲𝘆 𝗹𝗮𝘂𝗻𝗱𝗲𝗿𝗶𝗻𝗴 (𝗠𝗟) 𝗮𝗻𝗱 𝘁𝗲𝗿𝗿𝗼𝗿𝗶𝘀𝘁 𝗳𝗶𝗻𝗮𝗻𝗰𝗶𝗻𝗴 (𝗧𝗙) 𝗿𝗶𝘀𝗸𝘀 across the EU's financial sector from 2022 to 2024. The EBA, mandated to issue such an opinion biennially, identifies evolving threats driven by technological innovation, including vulnerabilities in FinTech, RegTech, and crypto assets, alongside the 𝗶𝗻𝗰𝗿𝗲𝗮𝘀𝗶𝗻𝗴 𝘀𝗼𝗽𝗵𝗶𝘀𝘁𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗼𝗳 𝗳𝗿𝗮𝘂𝗱 𝗮𝗻𝗱 𝗰𝘆𝗯𝗲𝗿𝗰𝗿𝗶𝗺𝗲 𝘀𝗰𝗵𝗲𝗺𝗲𝘀. While acknowledging positive developments like reduced tax crime risks and improved supervisory engagement in certain areas, the EBA highlights persistent challenges such as 𝗶𝗻𝗰𝗼𝗻𝘀𝗶𝘀𝘁𝗲𝗻𝘁 𝗮𝗻𝘁𝗶-𝗺𝗼𝗻𝗲𝘆 𝗹𝗮𝘂𝗻𝗱𝗲𝗿𝗶𝗻𝗴 𝗮𝗻𝗱 𝗰𝗼𝘂𝗻𝘁𝗲𝗿-𝘁𝗲𝗿𝗿𝗼𝗿𝗶𝘀𝘁 𝗳𝗶𝗻𝗮𝗻𝗰𝗶𝗻𝗴 (𝗔𝗠𝗟/𝗖𝗙𝗧) 𝘀𝘆𝘀𝘁𝗲𝗺 𝗲𝗳𝗳𝗲𝗰𝘁𝗶𝘃𝗲𝗻𝗲𝘀𝘀 𝗮𝗻𝗱 𝘁𝗵𝗲 𝗰𝗼𝗻𝘁𝗶𝗻𝘂𝗲𝗱 𝗽𝗿𝗼𝗺𝗶𝗻𝗲𝗻𝗰𝗲 𝗼𝗳 𝗰𝘂𝘀𝘁𝗼𝗺𝗲𝗿 𝗱𝘂𝗲 𝗱𝗶𝗹𝗶𝗴𝗲𝗻𝗰𝗲 (𝗖𝗗𝗗) 𝘀𝗵𝗼𝗿𝘁𝗰𝗼𝗺𝗶𝗻𝗴𝘀. The report underscores the critical need for regulatory clarity and a more unified application of risk-based approaches throughout the EU's financial landscape.

This study develops a machine learning framework to identify high-risk enterprise financial reports, comparing Support Vector Machine, Random Forest, and K-Nearest Neighbors models. Using 2020–2025 audit data from the Big Four firms, Random Forest showed the highest performance (F1-score: 0.9012), excelling in detecting fraud and compliance issues. While KNN struggled with high-dimensional data, SVM performed well but was computationally intensive. The study highlights the potential of machine learning in auditing but notes limitations, including reliance on structured data and exclusion of external economic factors.

This paper tackles corporate fraud detection using real-world Chinese stock market data. It highlights challenges like information overload and hidden fraud. The proposed KeGCNR model enhances detection with knowledge graph embeddings and robust training. Experiments show superior performance. Future research should address class imbalance and IND noise. Public datasets are provided.

"the typical organization loses 5% of revenues yearly because of fraud. Businesses are subject to fraud risk, and it is critical for organizations to put in place effective control mechanisms to prevent fraud".

"#frauddetection is overwhelmingly associated with the greater field of #anomalydetection, which is usually performed via unsupervised learning techniques because of the lack of labeled data needed for #supervisedlearning. However, a small quantity of labeled data does often exist. This research article aims to evaluate the efficacy of a deep semi-supervised anomaly detection technique, called Deep SAD, for detecting #fraud in high-frequency #financialdata."

Amid #digitalfinance's rise, its role in combating corporate #financialfraud gains attention. The study explores how digital finance curbs fraud via #transparency, #regulation, #riskcontrol, and trust mechanisms. Findings suggest positive impacts on deterring corporate #fraud, with implications for digital finance development and #fraudprevention

#financialrisk #prediction is vital but hindered by outdated algorithms and the absence of comprehensive benchmarks. Addressing this, FinPT uses large pretrained models and Profile Tuning for #risk prediction, while FinBench provides datasets on #default, #fraud, and #churn. FinPT inserts tabular data into templates, generates customer profiles using #languagemodels, and fine-tunes models for predictions, demonstrated effectively through experiments on FinBench, enhancing understanding of language models in financial risk.

The article discusses the use of #deeplearning and #datamining in business intelligence protocols to optimize data-driven decision-making and improve efficiency. The authors focus on the use of Graph Neural Network and Autoencoders Models to process large amounts of data and model #fraud behaviors. They suggest that deep learning can be used to control #moneylaundering in financial institutions and improve visibility and transparency in businesses.

Local communities exposed to #fraudulent #investmentadvisory firms tend to withdraw deposits from their affiliated #banks, even though the banks are not involved in the #misconduct. The #reputationalrisk is more significant when banks share names with fraudulent advisory firms or are located in areas with high social norms. The author establishes causality by exploring a quasi-natural experiment in which #fraud is likely exogenously revealed.