|
|
Matthias Knab, Opalesque: Key points
- New FSB report advises authorities how to monitor AI adoption and related risks in finance.
- Proposes direct and proxy indicators to track AI use and emerging vulnerabilities.
- Case study flags concentration risks in the GenAI supply chain (hardware, cloud, models) and dependence on a few critical third parties.
The Financial Stability Board (FSB) has published guidance on how authorities can better monitor the adoption of artificial intelligence (AI) in the financial sector and the vulnerabilities that may arise. The paper builds on the FSB's 2024 work on the financial-stability implications of AI and notes that supervisory monitoring remains at an early stage due to data gaps and the lack of standardised taxonomies.
What the FSB recommends
- Develop and align common taxonomies for AI use cases across authorities.
- Use a mix of direct indicators (e.g., number of AI models in production, model classes in use, AI spend) and proxy indicators (e.g., cloud/compute intensity, data-pipeline reliance) to assess adoption and risk.
- Track vulnerabilities stemming from AI, including model risk, operational risk, data quality, bias/fairness, cyber exposure, and contagion via common third parties.
Third-party and GenAI concentration risks
The report highlights growing reliance on a small set of AI suppliers across the GenAI stack - specialised chips, cloud infrastruct...................... To view our full article Click here
|
|
RSS
