RCVS roundtable event sets the ball rolling on guidance for AI use in the veterinary sector

How the veterinary professions might flourish in the midst of exponential technological change was discussed earlier this week, as we hosted our artificial intelligence (AI) roundtable to consider how AI use in the veterinary sector might best be regulated to optimise positive outcomes and minimise risk.

The event took place on Monday 20 May 2024 at the Institution of Engineering and Technology (IET) in London, starting with an introduction by RCVS CEO Lizzie Lockett in which she gauged how the delegates were feeling about AI, and outlined the principles and aims of the day.

This was followed by a series of introductory talks outlining not only key concepts in AI, but how it was currently being used in the veterinary sector and how it might develop in the future.

Dr Liz Barton MRCVS, Head of Communications at teleconsulting company Vet.CT, spoke about its application in clinical practice including in preventative medicine, diagnostics, treatment and prognostics, for example, in predicting likely surgical outcomes and so improving risk management and patient outcomes. In particular, she highlighted that the use of AI tools in clinical practice for tasks such as pattern recognition had led to many unforeseen and unexpected benefits, for example, by picking up things that humans may not.

Dr P-J Noble, Senior Lecturer in Small Animal Science at the University of Liverpool, spoke about the use of AI and data in the veterinary sector, particularly highlighting its use in the university’s Small Animal Veterinary Surveillance Network (SAVSNET). As a project that needs to use large quantities of data from veterinary practices in order to be effective, Dr Noble said that AI tools had proven useful in processing, assessing and annotating qualitative data such as clinical notes once they were programmed to recognise certain patterns, saving lots of research time and costs. He emphasised the importance of making AI tools ethical and explainable – meaning the connection between the data inputted into the tool, and the outputs were clear.

Dr Chris Trace MRCVS, Head of Digital Learning at the University of Surrey, spoke about AI use in higher education and how it has already started to be used beneficially both as a teaching and learning aid, as well as for assessment and feedback. He highlighted tools that had been developed to enhance veterinary education including virtual reality programmes, which allowed students to undertake interactive veterinary work in a variety of virtual reality (VR) scenarios, based on good clinical data. On the principles of AI, he reminded delegates that, today, AI technology is the least capable it will ever be, and that the pace of change means it will soon be able to undertake more tasks.

The final introductory talk was from Joanne Conway, a Deloitte Partner in AI & Digital Regulation, who spoke about the current global regulatory climate for AI and how the risks were being managed. She said that though the regulatory environment for AI was patchy, there was already a number of existing laws in the UK and other jurisdictions that were relevant – for example, the General Data Protection Regulation as well as intellectual property rights and protections. In terms of the ethical framework she said the European Union’s AI Act was likely to have international implications as the first piece of legislation of its kind and that it took a risk-based approach to AI, taking into consideration impacts such as job displacement.

Following the introductory talks, RCVS Registrar, Eleanor Ferguson, asked delegates to work in groups to discuss and illustrate what they thought the veterinary world would look like if AI worked perfectly for the veterinary world, and the potential ‘doom’ scenario if it did not.

The afternoon session was led by the Darren Tysoe, RCVS Chief Technology Officer, in which groups of delegates were asked to discuss a number of different practical questions over AI use in the veterinary professions and how it might be regulated. For example, there were discussions on how to help vet and VN students make the most of AI in learning and assessment, responsible use of AI in clinical settings, the risks of not using AI, how veterinary professionals can work with animal owners and keepers to ensure the safe and productive use of AI, and whether AI-led devices should be regulated.

Following the discussions Junior Vice-President Linda Belton encouraged delegates to feed back their answers to these key questions, all of which will be included in a formal report of the day published this summer.

Commenting on the day Lizzie Lockett said: “This was a really positive and exciting event that involved a heady mix of trepidation over the risks and implications of AI now and into the future, and optimism over the beneficial impact it could have for education, diagnosis, treatment and patient outcomes. The discussions have certainly given us at the College a lot of food for thought on how we can put in place guardrails and guidelines on the appropriate use of AI in the veterinary sector.

“This is an area of technology that is evolving so rapidly that it would not be effective for us to put in place specific guidance for the use of AI tools, but instead we will be looking at how we can regulate the use of AI in the round and ensure our principles are sufficiently future-proofed to keep up with the pace of change.

“Any regulation will start with first principles, such as transparency and honesty around the use of AI in veterinary practice, the minimisation of potential risks, and the continuing importance of professional accountability for decision-making, even where such decisions may have been heavily influenced by the use of AI tools and AI-generated data.”

The input gathered from the roundtable will now be considered by a range of RCVS committees over the coming months. A full report of the event will be published this summer.