Approaching AI in the legal sector: Law Society response to Government AI whitepaper

As organisations and individuals consider how best to leverage AI and establish appropriate governance guardrails in their own contexts, it’s important to retain sight of the overall picture and ensure that industry regulation is robust and proportionate.

We were pleased to work with the Law Society on their response to the UK Government’s whitepaper A pro-innovation approach to AI regulation.

Our key takeaways are as follows.

1. Whilst we are at an early stage, AI offers significant potential to transform the legal profession, including to improve efficiency and to democratise access to legal services.
2. Allowing industry regulators to adopt responses specific to their areas of responsibility is desirable; combining sector-specific regulation with firm legislation focusing on high-risk contexts and dangerous capabilities, with appropriate definitions and guidance, would provide a more comprehensive safety net.
3. It is important to be clear about how discrepancies will be dealt with across regulatory regimes; regulators may need additional support, guidance and resources to implement the Government’s proposed principles-based approach.
4. Clarity is needed where legal services are provided across jurisdictions whose regulatory regimes may differ significantly; definitions should harmonise with international benchmarks such as the OECD definition or those set out in the National Security and Investment Act to avoid conflict, reduce uncertainty and ensure those regimes are interoperable.
5. Urgent regulation is required to clarify when and how entities may be liable for the outputs of an AI system.
6. Client confidentiality is critical and must be a key component of future regulation.
7. As well as requiring that AI outputs are transparent and explainable, it is important that those outputs can be readily interpreted so that errors in the logic or decision-making can be located and corrected.
8. Procurement processes should be risk-sensitive, considering the benefits, risk and limitations of AI holistically.
9. Notwithstanding the potential, human judgement remains crucial and those operating AI systems should ultimately retain responsibility and accountability; understanding client needs, level of risk and vulnerability, and providing strategic, commercially-focussed advice remain the domain of human, legally qualified professionals.
10. In summary, nuance and balance is key to ensure that AI systems can be developed creatively and leveraged usefully by the legal profession, whilst maintaining appropriate clarity and protection for clients and practitioners.