More Opportunity, Less Risk: 8 Steps to Protect Financial Services Data with GenAI
Use GenAI more effectively with these fundamentals to select, store, and access data
Executive Summary
Generative Artificial Intelligence (GenAI) offers financial services institutions enormous opportunities, particularly in unstructured dataset analysis and management. It may also add to and/or exacerbate security risks.
Understanding and mitigating those risks may be a fundamental shift for firms looking to leverage GenAI capabilities. For that reason, the FS-ISAC Artificial Intelligence Risk Working Group researched data governance and developed eight foundational steps to use GenAI effectively and cautiously as you select, store, and access data.
Those steps include:
- Risk identification
- Data selection
- Data lineage
- Data access
- Customer privacy
- Test plans
- Model vulnerabilities
- Vendor data storage
This paper aims to help financial institutions make decisions about data governance aligned with their GenAI usage, budget, and risk appetite by providing the foundational governance steps necessary for safer, more effective data governance and routine refinements in the age of GenAI. (An upcoming paper will formalize a GenAI data governance framework and prioritizations.) Our intention is to help you build a solid foundation of clean data, accurate inventory, explainable policies, continued vigilance, and discipline in protecting sensitive datasets.
With that information, financial services firms of any size are better able to develop an approach to data governance that harnesses the benefits of GenAI and remains controllable, compliant, and ethical.