Vicky Crichton: Effective oversight of AI essential for public acceptance and trust

How confident would any of us feel in identifying when decisions might have been made about us using AI? How comfortable are we with that?

For the legal sector, AI offers clear potential benefits to both clients and lawyers. It could open up new legal advice and support services for consumers, leading to improved access, choice and cost. Inequalities could be reduced through smarter translation support. For lawyers and firms, bringing AI to tasks like document management could reduce costs and improve due diligence. This could leave lawyers more time to focus on client contact, legal reasoning, negotiation, and all the parts of the role where human intelligence and human contact is vital.

In its consultation on an AI strategy for Scotland, the Scottish Government asks how AI could benefit Scotland’s people. We think the legal sector has a good and ever-developing story to tell here.

Public benefit isn’t the only part of the puzzle. The consultation also asks how public confidence can be built in AI as a trusted, responsible and ethical tool. That’s a tougher challenge.

The public will expect and believe there is some effective oversight of AI, and its application in the services and products that they use. Confidence in this oversight – or regulation – will be a key part of believing AI can be trusted, responsible and ethical.

This isn’t specific to legal services. Regulators across all sectors will be grappling with the challenge of the many aspects of AI that may already fall within the scope of their work. This is the case wherever entities or individual professionals are providing services which could be using AI to support work or decision making.

Nesta, the innovation foundation, has suggested a move to ‘anticipatory regulation’ would better support and regulate disruptive innovation, allowing regulators to be quicker to respond to new developments.

It’s vital that regulators have the tools and skills to play a positive role in creating the ecosystem for AI to flourish. If not, then there’s a greater risk of them becoming a barrier to innovation, or failing to address a key risk to the public.

Key potential harms do need to be considered and prevented. For example, bias in decision-making impacting unfairly on certain groups could have a significant impact on both fairness and public confidence in AI. Of course, human decisions are also obviously subject to bias, and it’s important to ensure we don’t simply replicate this in the AI we develop, but seize the opportunity to reduce bias.

At the SLCC, we could find ourselves at the frontline of consumer dissatisfaction arising from the use of AI in legal services, and having to consider how to approach those complaints. There might also be real opportunities for our work, for example in helping us spot further trends and patterns in complaints which could help manage risk or identify improvements. With reform of legal regulation on the horizon in Scotland, there’s an opportunity to future proof it to meet the challenge of disruptive innovations like AI. To recognise the huge benefits that developments like AI could have for lawyers and consumers alike. To create an infrastructure that supports and rewards that innovation, while providing public confidence that its risks are being monitored and managed. That helps us take the leap into the unknown future armed with the tools we need to deal with it.

Vicky Crichton is Director of Public Policy, Scottish Legal Complaints Commission