FINOS Introduces AI Adoption Framework
- Fintech Open Source Foundation (FINOS) has published the first draft of an AI adoption framework
- This is good news for an industry eager to implement LLMs but facing significant obstacles
- The FINOS framework outlines 16 control procedures to limit 14 specific threats
We recently noted that 72% of finance leaders actively use AI in their operations, showing that FI leaders see the importance of AI technology. However, unlike the EU, the USA has been slow to adopt regulations for AI, particularly large language models (LMM).
Before partnering with new technology vendors who deploy AI technologies -- such as OrboGraph and our OrbNet AI and OrbNet Forensic AI -- many banks are performing an AI Model Validation. However, the specifics vary from FI-to-FI, and the need for consistency is clear.
It appears that the Fintech Open Source Foundation (FINOS) -- an umbrella organization under the Linux Foundation, whose purpose is to accelerate collaboration and innovation in financial services through the adoption of open source software, standards and best practices -- is taking the first step to help regulate the industry.
Risks and Governance for AI Adoption Framework in Banking
As reported by Banking Dive, The Fintech Open Source Foundation (FINOS) has taken an important step in guiding the financial industry's adoption of large language model (LLM) technologies, having published the first draft of an AI adoption framework that details common risks and governance measures to help banks and other firms successfully deploy these powerful AI tools.
FINOS is comprised of an impressive roster of participating Fintechs, cloud providers, and FIs, including:
- Nvidia
- Credit rating agency Moody
- Capital One
- Citi
- Goldman Sachs
- JPMorgan Chase
- Morgan Stanley
- Amazon Web Services
- Microsoft
- Google Cloud
- ... and many more...
This framework is a critical development, as the financial sector has been eager to implement LLMs but has faced significant obstacles. The "main obstacle to adoption of GenAI in financial services is the lack of a compliance framework," according to FINOS Executive Director Gabriele Columbro. With regulators and lawmakers still grappling with how to approach the risks of generative AI, this industry-led effort provides much-needed guidance.
“This is a great starting point for us, as an industry, to collaborate on a structured approach to the adoption and governance of AI,” Madhu Coimbatore, head of AI development platforms at Morgan Stanley, said in the Tuesday announcement.
The FINOS framework outlines 16 control procedures to limit 14 specific threats, including data leakage, ineffective encryption, and model hallucinations. This structured approach to AI readiness and governance will be invaluable for financial institutions looking to harness the benefits of LLMs while mitigating associated risks.
Risks and Governance for AI Adoption Framework in Banking
As noted by White & Case LLP:
Currently, there is no comprehensive federal legislation or regulations in the US that regulate the development of AI or specifically prohibit or restrict their use. However, there are existing federal laws that concern AI albeit with limited application.
Banking Dive notes that the EU is already further along with regulating AI, with the European Union AI Act taking effect this past August. So, will the USA take notice and utilize this a roadmap for their regulations?
One thing is for certain: Regulations are a necessity for AI, as its application and potential do not necessarily outweigh the risk of unchecked deployment.