Minutes of the London FXJSC Legal Sub-committee Meeting – 18 June 2024

Minutes

Minute 1: Introduction

The Chair (Sharon Blackman) opened the meeting.

Minute 2: Minutes

The minutes of the meeting of 5 March 2024 were approved.

Minute 3: AI and Financial Services update from Latham & Watkins (Becky Critchley & Gabriel Lakeman)

The speakers presented on the following five areas within AI.

Overview of AI regulation in the UK

There is no single overarching legislative framework or single regulatory function to govern AI. Instead, there is a non-statutory, principles-based approach to be implemented by sector-specific regulators.

There are a variety of examples of the rules and legislation that already apply to the use of AI in a financial services context, including:

• Consumer Duty: The Financial Conduct Authority's Consumer Duty imposes broad "crosscutting" obligations on firms, including the requirement to "act to deliver good outcomes for retail customers". The key issue is how to ensure use of AI is compliant with the Consumer Duty, including ability to foresee harm
• Outsourcing/Third-party risk management ("TPRM"): Firms are subject to a detailed regime when outsourcing critical or important functions, but the key issue is the extent to which the use of AI involves outsourcing/ reliance on third-party providers, and has been assessed against the firm's outsourcing policies and procedures.
• Senior Manager & Conduct Rules ("SM&CR"): SM&CR aims to reduce harm to consumers and strengthen market integrity by making individuals more accountable for their conduct and competence. The key issue is who should ultimately be responsible for AI, including under the SM&CR.

Developments in AI use

The speakers noted that they are being asked increasingly about AI assistance, although they haven't yet had clients ask for them to use co-pilot in meetings. The speakers outlined recent examples where AI is being used and in what context, highlighting that this is an area to watch.

Key risk considerations

There are a variety of emerging AI risks, some were highlighted:

• Making unsupported recommendations
• Lack of human oversight/interaction
• Conflicts of interests
• Completely false information (hallucinations)
• Bias and discrimination
• Inadvertent performance of regulated activities
• Data privacy and security
• Market stability

There are Risk Mitigation Strategies that can be adopted by firms, to ensure alignment with the current regulatory obligations:

• Firms can understand how legal and regulatory obligations interact with AI use, and ensure that they are able to explain their use of AI to regulators.
• Carry out risk assessments and assign risk ratings
• Consider and account for data protection considerations
• Consider governance arrangements and that a robust AI governance framework is in place
• Consider the existing systems and controls and policy framework and whether any enhancements should be made

The Global FX Code and AI principles

The speakers explained that there six leading principles under the Code which are relevant in the context of AI and under which the use of AI can be considered in an FX context:

• Ethics: Market Participants are expected to behave in an ethical and professional manner to promote the fairness and integrity of the FX Market.
• Governance: Market Participants are expected to have a sound and effective governance framework to provide for clear responsibility for and comprehensive oversight of their FX Market activity and to promote responsible engagement in the FX Market.
• Execution: Market Participants are expected to exercise care when negotiating and executing transactions in order to promote a robust, fair, open, liquid, and appropriately transparent FX Market.
• Information Sharing: Market Participants are expected to be clear and accurate in their communications and to protect confidential information to promote effective communication that supports a robust, fair, open, liquid, and appropriately transparent FX Market.
• Risk Management and Compliance: Market Participants are expected to promote and maintain a robust control and compliance environment to effectively identify, manage, and report on the risks associated with their engagement in the FX Market.
• Confirmation and Settlement Processes: Market Participants are expected to put in place robust, efficient, transparent, and risk-mitigating post-trade processes to promote the predictable, smooth, and timely settlement of transactions in the FX Market.

Emerging ethical issues

There are emerging ethical issues that firms are grappling with when developing and utilising AI. Comparisons were drawn between approaches in the EU, UK and other jurisdictions. The speakers noted that the EU approach is quite interesting as they have come in with a legislative regime very early on.

The use of AI in the supply chain was discussed with the participants asking how far they should go to ensure that their suppliers are managing AI. The speakers noted that this is a question that could be asked of suppliers so that they declare what AI they are using, and firms can consider this in their AI risk management frameworks.

Minute 4: The Post Office Inquiry and ethics in the legal profession (Jonathan Ashley-Norman KC, Three Raymond Buildings Barristers)

Mr Ashley-Norman started by giving an historical overview of the events that led to the Health & Safety at Work Act ("HSAW") 1974, which was developed in response to the particularly dangerous employment conditions that existed the UK at that time. He explained that, amongst other things, the HSAW 1974 made it a criminal offence for an employer to conduct its undertaking in such a way that it fails to ensure the safety of its employees. This created an extremely wide duty on the boards of companies, specifically the individual responsible for health & safety. As risk assessment regimes and codes of practice have been finessed over the years, there has been a corresponding decrease of deaths in the health & safety context. The HSAW 1974 has in large measure completed its task by instilling in the board room the matters relating to health & safety.

Other legislation can be looked at as doing effectively the same thing in instilling boardroom responsibility in other areas, by acting as the deterrent and changing corporate behaviour, e.g., Proceeds of Crime Act 2002, Bribery Act 2010 and Economic Crime Act 2023.

In relation to the Post Office, it will have had managers and directors subject to such obligations, various issues were brought to their attention, and yet they seemingly failed in their duty of care, and so will potentially be at the wrong end of criminal investigation.

Mr Ashley-Norman then put this into the context of the Global FX Code, which is organised around six leading principles: Ethics, Governance, Execution, Information Sharing, Risk Management and Compliance and Confirmation and Settlement Processes. The history of dishonesty, the Ghosh test, which is an objective test, was discussed alongside the Ivey test.

It was noted that an organisation's culture is hard to change and influence.

Minute 5: Any other business

There was no other business discussed.

Attendees

Sharon Blackman (Chair) - Citigroup
Baljit Saini - NatWest (virtual)
David Harris - Financial Conduct Authority (virtual)
Rakesh Shah - Standard Chartered (virtual)
Rowland Stacey - Goldman Sachs
Simon Goldsworthy - Deutsche Bank (virtual)
Tamsin Rolls - JP Morgan Chase

FXJSC Legal Sub-Committee Secretariat

Sakshi Gupta - Bank of England
Matthew Hartley - Bank of England
Carly Jones - Bank of England

Guest attendees

Becky Critchley - Latham & Watkins
Gabriel Lakeman - Latham & Watkins
Charlotte Branfield - Three Raymond Buildings Barristers
Jonathan Ashley-Norman KC - Three Raymond Buildings Barristers

Apologies

Gaynor Wood - CLS
Harkamal Singh Atwal - HSBC
Krisha Somaiya - UBS
Mayank Patel - Bank of America
Nimisha Kanabar - Morgan Stanley
Sunil Samani - XTX Markets