XBRL (eXtensible Business Reporting Language) US has joined the Center for Research toward Advancing Financial Technologies (CRAFT) as an affiliate partner to increase the efficiency and accuracy of financial information fed to large language models (LLMs).
XBRL is the global open standard for digital business reporting. The language transforms human-written financial documents into machine-readable structured data by assigning digital tags to financial and business information. XBRL uses taxonomies defined by regulators to categorize and structure data. The standard is mandated by multiple organizations, including the Securities and Exchange Commission (SEC), Federal Energy Regulatory Commission (FERC), and the Federal Deposit Insurance Corporation (FDIC). XBRL improves readability and searchability while reducing human error when digitizing financial records.
CRAFT is an Industry University Cooperative Research organization, funded by the National Science Foundation. Co-founded by Stevens Institute of Technology and Rensselaer Polytechnic Institute, CRAFT aims to bridge the gap between traditional finance and fintech by developing research and policies to ensure the efficacy and safety of entities transitioning into the industry. The program brings academics and industry professionals together to “secure our financial data, create and test more equitable trading platforms, [and] inform financial regulations.” Along with collaborative fintech research, CRAFT builds tools for risk management and stress testing to further address the challenges in the financial sector.
LLMs use XBRL’s data to provide users with information and financial analysis. With CRAFT, XBRL will help auditing companies improve the efficiency, quality, and accuracy of research, particularly within decentralized finance (DeFi). Decentralized finance is a noncustodial ecosystem in which there is no intermediary, enabling peer-to-peer lending. However, DeFi comes with a high financial risk, including theft, volatility, and potential loss of investment.
Looking forward, with LLMs’ exponential growth in the financial sector, the accuracy of data fed through software becomes increasingly important. The data AI systems are based on are directly linked to the reliability and ethical use of said systems. Standardized and regulator-backed data like XBRL decreases the chances of biased, misleading, or hallucinatory data outputs. AI systems will be grounded in consistent and reliable information and policies. Ultimately, the partnership will expand the role of financial data and algorithms, increasing the reliability and transparency of preexisting systems across traditional and decentralized markets.
