AI matters | Financial services round-up (September 2024)

ICMA responds to EU Commission's consultation on AI in the financial sector

Friday the 13th may be associated with unluckiness, but in the context of AI in Financial Services, Friday 13 September 2024 was a particularly propitious day as it saw the publication by the International Capital Markets Association (ICMA) of its response to the EU Commission's consultation on AI in the financial sector, which we covered in our June 2024 round-up.    

The response draws on feedback from a subset of ICMA's AI in Capital Markets working group, comprising investors, banks, market infrastructures, professional services firms and issuers across the international debt capital markets.

Here are the key takeaways:

• Members of ICMA encourage innovation across the debt capital markets industry and support fair and transparent applications of AI.
• In general they have had a positive experience when using AI, especially its ability to increase operational efficiency and free capacity for people to focus on more high value tasks.
• The responsible use of AI in an organisation requires human validation, a robust internal governance framework, staff education and training, and guidelines.  Internal risk frameworks that address AI-related risk are already in place within financial organisations. 
• Members emphasised the strong ties between AI technology and developments related to ESG and DLT and the risk of innovation being impacted if supplementary regulation to the AI Act is brought in too soon.
• In implementation the EU AI Act, legislators should take into account the existing EU financial services dossiers such as UCITS, AIFMD and MiFID II/MiFIR, which already require safeguards to be in place for the responsible use of technology. 

Supervision Innovators Conference 2024

While possibly not everyone's go-to annual jamboree, last week's Supervision Innovators Conference 2024 in Frankfurt, underscored the fundamental role that AI plays within supervisory technology tools (SupTech).  

Elizabeth McCaul, a member of the European Central Bank's supervisory board noted that “[w]e are committed to delivering additional breakthrough solutions that use AI – and more specifically generative AI – to simplify and automate workflows, while improving collaboration within European banking supervision.”  She drew attention to the ongoing use of tools such as AthenaGPT,  which enables supervisors to interact with several supervisory information sources at the same time.  McCaul's comments were in the context of the overall supervisory strategy of banking supervision in Europe, which is centred on people and technology.

New Bank of England AI Consortium

On 25 September 2024, the Bank of England announced that it is establishing an AI Consortium, bringing together stakeholders from the public and private sector stakeholders to:

• identify existing and potential new uses of AI in financial services
• to discuss the benefits, risks and challenges that arise from the use of AI
• to input into the Bank of England’s approach to how it addresses risks and challenges, and promotes the safe adoption of AI.

Applications to join the Consortium must be submitted by 8 November 2024.

Bank of England speech: TRUSTED AI and the human touch 

One of the unintended beneficiaries of AI has perhaps been conference organisers!  Speaking at at the inaugural Central Bank AI Conference, James Benford, Executive Director for Data and Analytics Transformation and Chief Data Officer at the Bank of England, reflected on the opportunities and challenges arising out of the growth in AI solutions in the context of the Bank of England's role in delivering monetary and financial stability.  

Benford summarises some of the work the Bank has done on modelling over the years.  The safe and effective development and adoption of AI models by the Bank requires that they are TRUSTED (Targeted, Reliable, Understood, Secure, stress-Tested, Ethical and Durable) and informed at all times by three human attributes: confidence, resolve and humility.  One final point worth noting from his speech, which has application to all organisations grappling with the AI revolution: the need for AI fluency for all of the organisation's expert data professionals and AI literacy for staff generally.