The ICAEW Annual Conference took place at the same time that the UK government’s AI Safety Summit at Bletchley Park reconvened for its second day (2 November). While this was coincidental, the two events shared a theme, as ICAEW’s sessions covered a huge breadth of technology-related themes including cloud adoption, ESG, digital transformation, automation and the all-important people side of technology. But if there was one topic that dominated the conversation, it had to be artificial intelligence (AI).

AI, work and the era of the copilot

In the Conference’s first keynote speech, Microsoft’s Nir Evron explored why, rather than replacing humans, AI has the potential to help us tackle the challenges of ‘digital debt’. This was a sentiment echoed later in the day by FreeAgent’s Tony Stevenson, who touched on the ‘lump of labour fallacy’ (the misconception that there is a finite amount of work to be done). Indeed, Craig Clarke, Director of Product at FreeAgent, said: “The promise of AI is that it will lift the burden of mundane, repetitive activity away from us to leave more capacity, context and care to tackle the challenges only the human condition can master.”

Evron emphasised the importance of responsible AI, alongside appropriate data controls, before explaining ‘the era of the copilot’. An AI copilot trained on the UK’s Financial Reporting Standards was demonstrated, taking natural language prompts and translating hundreds of pages of complex, dense standards into straightforward, referenced responses that could be interrogated.

Another copilot, from Engine B, showed how accountants could work side by side with AI models specialised at extracting and analysing data to inform the others’ thinking. With the example of lease accounting, it was possible to show how a specially developed generative AI solution could be used to interrogate a client’s financial data and lease agreements to deliver answers to typical audit questions. Such an approach would be well suited to meet the reporting, audit and assurance demands of increasingly vast, complex and unstructured data sets.

AI challenges and opportunities

ESG reporting will require much greater automation and AI in order to process the enormous quantities of data it will create. These tools will need to be supported by accurate and timely data, precise frameworks and parameters, and intelligent interpretation to maximise the benefit. MHA experts flagged a key consideration around the cumulative emissions of AI models and that a role for our profession is to balance the benefits the tools can bring to audit and assurance of ESG reporting, with the emissions associated with developing the tools. 

In the keynote speech from Zoe Kleinman, BBC Technology Editor, we were told how AI is being used to significantly cut the level of emissions from air travel, including halving the ‘vapour trails’ that are a major contributor to air pollution.

Kleinman also highlighted other exciting developments driven by AI, including in neurotechnology, medicines and nuclear fusion. Kleinman asserted that we are most likely talking months, not years, for AI to begin transforming the way we work, in the same way that the motorcar transformed the streets of New York almost overnight in the early part of the 20th century.

AI also shows potential in how it can help businesses and firms navigate the many complexities of tax. Russell Gammon from Tax Systems explained that while many generative AI tools are not very good at maths and probably shouldn’t be trusted to complete tax returns just yet, they are good at tagging, mapping and consulting knowledge repositories. This could allow tax professionals to spend more time providing advice and less time on manual, repetitive transaction categorisation or poring over hundreds of pages of HMRC guidance.

Another crucial balancing act attendees were called to consider was that of encouraging the use of technology while listening to the concerns of colleagues. There’s also the risk of colleagues taking inappropriate risks in using open-source tools. 

Many noted that the technology would need to be useful to colleagues and employees for them to actually use a tool. For the same reasons, you cannot stop people from using open-source tools. If it is useful, people will use it. By educating colleagues on the risks and issues relating to these tools and providing less risky alternatives, potential issues should be mitigated. Many called on the importance of skills and training for the profession. Upskilling will be a lifelong journey.

ICAEW has created guidance on Generative AI for members.