Information on reviews into the use of AI (Artificial Intelligence) - May 2023

Have you since 01/01/2020 conducted any reviews into the use of AI? 

If so can you publish the findings along with methodology used?

Please note that we have broadly interpreted "AI" (Artificial Intelligence) to mean the simulation of human intelligence processes by machines. However, we note that there is currently no generally accepted definition of AI. 

By way of background, the FCA’s current internal usage of AI is limited to the conversion of speech to text (Natural Language Processing, NLP) and some limited project-specific proof-of-concept (PoC) experimentation.

The FCA is committed to using data and technology to make it a more effective and efficient regulator. To support this we have invested in data science capability and capacity across the organisation to understand how the use of data science and advanced analytics can be used to bring new insight.

We have completed a number of Natural Language Processing projects to allow for faster and more effective analysis of unstructured text documentation. This includes deploying a NLP platform which is now being used by a small number of users across the organisation to triage and prioritise document sets for review.

We have also delivered small scale PoCs which use some technologies including Deep Learning (e.g. for semantic search across text documentation); Object Recognition (e.g. for detecting images within documents); and Supervised Machine Learning (e.g. for helping to prioritise workloads). All of these projects are still at PoC and have not been scaled for use across the organisation.

You may also find the following FCA speeches of interest:

• Jessica Rusu’s speech dated 10 February 2022 (at our RegTech Forum webinar on the FCA’s approach to innovation and the growing importance of Supervisory Technologies)[1]
• Jessica Rusu’s speech dated 24 January 2023 (at The Alan Turing Institute’s Framework for Responsible Adoption of Artificial Intelligence in the Financial Services Industry (FAIR) event)[2]

Other helpful information

Details of our work[3] are regularly published and publicly available - please see the links below which also contain details of any methodologies used where applicable:

• FCA and Bank of England’s Artificial Intelligence Public Private Forum (AIPPF) Final Report (see here[4] and here[5])
• FCA and Bank of England report on findings and analysis of the 2022 machine learning in the UK financial services survey (see here[6] and here[7])
• FCA, PRA and Bank of England Artificial Intelligence Discussion Paper (see here[8] and here[9])
• FCA and Alan Turing Institute publication on AI in Financial Services[10]

The FCA was also involved in the following:

• IOSCO guidance for intermediaries and asset managers using Artificial Intelligence and Machine Learning[11]
• Digital Regulation Cooperation Forum Discussion Paper[12]

Please also see:

• Jessica Rusu’s speech dated 22 March 2022 (at our livestreamed session on the future of financial services with data and AI)[13]