GenAI in Action: Transforming Data Use in SupTech and RegTech

 

Delivered at the Singapore FinTech Festival, this session addresses the different ways GenAI can transform data use in SupTech and RegTech, as well as the best approaches to leverage this technology to improve supervision and regulation.

 

1. Real-world use cases of GenAI in SupTech and RegTech

Speakers touched on current implementations of GenAI in and outside their respective organisations.

  • Nino Landerer, Deutsche Bundesbank: A notable use case, RegBot (or SupBot), aids supervisors in analysing unstructured data, providing answers with sourced context to enhance decision-making.
    • Landerer also touched on the use of GenAI for climate-related financial risks, highlighting Project Gaia. This involved using GenAI to analyse large sets of unstructured data for climate risk indicators.
  • Patrick De Neef, De Nederlandsche Bank (DNB): DNB uses GenAI to enhance supervision by automating routine processes, freeing up time for creative, risk-identifying tasks.
  • Kenneth Gay, Monetary Authority of Singapore (MAS): GenAI enhances the work of central banks through avenues such as fast-moving event analysis, where it can automate scenario tracking and conduct predictive assessments during geopolitical or economic events.
    • MAS is already using GenAI for timely updates and risk evaluations, enabling staff to focus on high-level decision-making.
  • Other use cases raised by speakers include the use of GenAI to automate supervisory reviews as well as fraud and misconduct detection.

 

2. GenAI implementation

Susan Slocum of Reserve Bank of Australia (RBA) shared her experience of implementing ChatGPT, highlighting challenges and proactive steps taken to ensure its secure, ethical, and effective use.

  • Initial adoption challenges included security concerns and the need for controls. To address this, RBA adopted Australia’s AI Ethical Principles and the New South Wales AI Assurance Framework.
  • The RBA maintained a register of approved use cases to prevent repetitive queries and foster organisational learning.
  • Additionally, Slocum highlighted the importance of user education and empowerment. At RBA, they had introduced a prompt engineering course to help users maximise ChatGPT’s potential and concurrently educated employees on risks to foster responsible usage.

 

3. GenAI and ethical considerations

Slocum highlighted the importance of embedding ethical considerations into the use of GenAI in financial supervision. She outlined the risks of AI biases, hallucinations, and automation bias, emphasising the need for vigilance and human oversight.

  • Robust testing: In test runs of three use cases, it was found that current GenAI capabilities may not yet be reliable for all supervisory tasks.
  • Need for training and ethical vigilance: Supervisors and users must be trained to critically evaluate AI outputs.

 

As we embrace AI’s transformative potential, it is essential to balance innovation with ethical responsibility, structured implementation, and collaboration. By addressing biases, adhering to regulations, and fostering systems of diversified tools and expertise, we can unlock AI's capabilities while ensuring trust, inclusivity, and effectiveness across sectors.

 

Watch Full Session

 

 

Speakers:

  • Kenneth Gay, Executive Director, Enterprise Knowledge Department, Monetary Authority of Singapore
  • Nino Landerer, Deputy Director General Strategy & Innovation, Deutsche Bundesbank
  • Patrick De Neef, Chief Innovation Officer, De Nederlandsche Bank
  • Susan Slocum, Former Head of Innovation, Reserve Bank of Australia

 

Moderator:
  • Dr. Patrick Hoffmann, Advisor, BIS Innovation Hub - Singapore Centre
 
 

Join the GFTN Network