Advancements in technology happen at a rapid pace and financial authorities need to understand, adapt and embrace these changes. The NextGenAI Tech Showcase, hosted by Cambridge Suptech Lab with Alliance for Innovative Regulation (AIR), allowed financial authorities to understand the power and possibilities of GenAI, and explore solutions that could be leveraged for more efficient and effective consumer protection and supervision.
Various Gen AI-enabled solutions were demonstrated virtually, over three sessions. Each session catered to different themes aligned with the roundtable discussions on the advantages and concerns surrounding the use of Gen AI in the realm of financial regulation and supervision. Explore key learnings from the showcase here Key Learnings from NextGenAI Tech Showcase https://regulationinnovation.org/wp-content/uploads/2024/07/Key-Learnings-from-NextGenAI-Tech-Showcase.pdf
Sample Use Cases:
1. Data Analysis and Monitoring
- Generate synthetic data for data anonymisation.
- Interpret and summarise both structured and unstructured data. Unstructured data could include consumer complaints, in which natural language processing (NLP), topic modeling and sentiment analysis can be conducted to identify trends, patterns, and clusters.
- Extract insights from the extensive data held by financial regulators, including information from financial statements, regulatory compliance reports, consumer complaints, and corporate governance disclosures.
2. Risk Assessment
- Offer detailed risk assessment tools that enable financial regulators to derive crucial credit risk benchmarking metrics, such as the debt-to-equity ratio, non-performing loans ratio, and return on assets ratio, from the data stored in their repositories.
- Detect and interpret meaningful patterns inherent in the reporting data.
- Identify security flaws in financial systems by providing advanced capabilities for anomaly detection, testing, analysis, and continuous monitoring.
3. Compliance Automation
- Translate rules/regulations into supervisory processes and corresponding compliance forms/templates.
- Assist with coding tasks such as providing code completion suggestions, generating specific task-oriented code snippets, rectifying bugs, and code related documentation. This is intended to increase efficiency of the compliance process by ensuring standardised compliance patterns throughout the development lifecycle, generating detailed audit trails, and enabling quick adaptation to regulatory changes.
4. Consumer Protection
- Compare all available offerings for a particular financial product or service to identify the best option and serve as a consumer education tool.
- Provide context-sensitive language translation of financial information and multilingual support.
- Empower consumers’ financial literacy by providing assistance through virtual assistants or chatbots for various financial services-related procedures. A function of the virtual assistant could be to detect and respond in an appropriate language using conversational content from specified data sources, including links to websites, PDF documents, and other unstructured data.
5. Fraud Detection
- Provide insights into potentially deceptive activities. This could involve examining various indicators, such as an increase in complaints regarding hidden fees, lending rates exceeding stipulated terms and conditions, and a disproportionate rise in borrowings against collateral, among others.
6. Policy Development
- Generate regulatory documents and materials (content creation). For example, an application could involve promoting data-driven regulation by creating new policies or modifying existing ones. This could be achieved through extensive analysis of compliance data, resulting in better informed regulatory decisions.
7. Decision Support
- Interpret company disclosures and other digital reports.
- Prior to implementing, facilitate the evaluation of policies through policy simulation. In the context of banking, conduct back-testing to assess the impact of proposed policy changes, considering factors such as alterations in collateral threshold for borrowings, adjustments in lending to certain sectors or demographics, and changes in loan portfolio.
- Conduct analysis of extensive historical and current data to support the licensing process. As an example, a typical bank licensing process includes examining the applicant’s business plan, feasibility study, financial statements, compliance and operational history, as well as previous policies and practices. Generating insights from the aforementioned data provides financial regulators with a comprehensive understanding of the applicant’s track record.
- Facilitate the resolution of lawsuits for financial regulators. This could include analysing documentation, depositions, and reports that are in the form of unstructured data to provide useful insights. Additionally, it could also serve as an effective case management solution.
- Support financial regulators as a virtual assistant in retrieving information and generating reports.
