Contributors
Katie Simmonds, Managing Associate in our Digital team, recently joined a panel of experts for a webinar hosted by FStech. The webinar, titled 'Risk and reward: Navigating GenAI adoption in Financial Services' featured insightful discussions with Mark Katz, Client Strategy and Technology Officer for Financial Services at Hitachi Vantara and Prag Sharma, Global Head, Artificial Intelligence Centre of Excellence at Citi.
The webinar explored current use cases of GenAI by financial institutions and how they can build robust frameworks to harness GenAI’s potential while maintaining the highest standards of trust, compliance, and risk management.
Here are the key takeaways:
GenAI use cases
- Financial institutions are increasingly utilising GenAI applications across various functions including customer service, credit decision-making and fraud detection.
How to build trust-centred frameworks
- Ensure human oversight throughout the AI lifecycle, including training, interpretation of decisions, and future development to prevent bias and inaccuracies.
- Comply with data protection laws by adopting a 'privacy by design and default' approach, ensuring personal data is processed as expected and decisions are explainable.
- Focus AI use on low-risk processes such as information retrieval and KYC while using human validation for more complex tasks such as fraud detection and complaints handling.
- Consider potential biases in AI-driven risk assessments.
- Partner with specialist third-party providers of infrastructure and data-management.
- Collaborate with all key internal and external stakeholders from the outset.
- Be purposeful in how you decide to leverage AI.
- Keep up to speed with the evolution of the technology and remain aware of its limitations.
How to uphold transparency and explainability in AI decision-making
- Use high-quality data for training AI models, with significant human intervention to validate decision logic.
- Continuously fine-tune models and test outputs, particularly when using third-party AI providers.
- Design modular structures for AI to enhance reviewability and security.
- Explain to customers how processes work using understandable language.
If you missed the live session, you can watch the webinar recording here.
This article was also authored by Jai Thakrar, Trainee Solicitor at Womble Bond Dickinson.
This article is for general information only and reflects the position at the date of publication. It does not constitute legal advice.
