Allaying the AI anxieties

A recent study led by James Faulconbridge, a professor at Lancaster University, shows that accountancy firms have already moved through a number of stages to cope with AI – everything from highlighting the technology’s deficiencies to working with other professions (those with technology skills) to offer a remodelled service.

Faulconbridge and his team conclude that their study shows ‘more complex outcomes’ than the current narrative that depicts an apocalyptic ‘end of the profession’. ‘The challenge is finding a way to respond in a way that isn’t bad for your wellbeing,’ he says.

Value of training

That may seem a daunting prospect at the moment, as research by ACCA⁠ bears out. ACCA’s poll of 10,000 professionals across 157 countries found significant numbers feel ‘overwhelmed’ by the pace of technological change, particularly those under the age of 26, almost half of whom are anxious about it. Respondents were near-unanimous in wanting their employers to provide more training.

Further ACCA research shows that 75% of employers value an individual’s ‘willingness’ to learn new skills above their existing skillset.

Conversations across the profession highlight widespread worries about access to training and whether new skills can be acquired, with middle managers losing sleep over training their staff while hitting targets, and seniors brooding over the investment needed for AI systems and sourcing the training required. Nearly everyone worries they simply do not understand the technology or have the vocabulary needed to talk about its capabilities, limitations and risks.

All are agreed on one thing, though: it is training, engagement and familiarity with the technology that ultimately manages anxiety and wellbeing.

Personal roadmaps

Andrew Chong, founder of fintech firm uPledge and former AI expert with EY, stresses the need for firms to make time available for training, which should be mandatory. The difficulty that firms face here is that most have a ‘chargeable hours’ culture, which creates obstacles to training that is flexible or, in practice, voluntary.

In Chong’s experience, training is widely disliked if it is not well targeted. ‘Training is popular,’ he says, ‘if the relevance is high.’ He adds that it was only when he moved to a department where ‘100% utilisation’ was no longer a way to assess staff performance that he witnessed a real attitude change to training.

He does, however, stress that individuals will, to some extent, need to take charge of their own destinies and recommends people draw up their own personal ‘roadmaps’ for skills training.

Getting started

Cara Molyneux, a researcher with the Technology in Professional Services project at Lancaster University, says anxieties about the implementation of AI have focused on fears among managers about how to get started.

Making a start, as simple as it seems, is an important step to ease anxieties. Molyneux says employee engagement is needed at the earliest possible stage when considering how and where AI will be deployed. Then firms can aim for a ‘easy win’, which will ‘motivate’ staff.

‘It’s about starting slowly and not trying to do everything at once,’ she says. ‘Resistance to change should be eased and smoothed if you can show that AI is actually going to help in your day-to-day working life.’

Employer responsibility

There is no escaping the employer’s role in allaying fears. Lauren Branston, chief executive of the UK Institute of Business Ethics (IBS), says organisations need to realise that general fears and speculation are feeding into workplace concerns.

‘There’s a huge opportunity ahead,’ she says. ‘We do want it to work for people, but we do need a human-centric approach to this, and we also need one that is cognisant of the fact that people’s experience is not just in work, it is also outside of work, and their trust in technology is being challenged on many fronts.’

Branston believes ethics has a fundamental role to play at this transitional moment to guide employers in addressing anxieties and wellbeing. Employees need to be consulted about their concerns, those concerns need to be acted on, and the best ways found to solve AI workforce issues.

For Branston, employers should be ‘thinking about how we support people in taking this forward and feeling like they have agency to raise concerns that are acted upon’.

Not so smart

Perhaps the most important observation when it comes to wellbeing is that we have moved on from the shock launch of ChatGPT in late 2022, which triggered dystopian predictions of mass unemployment.

It is becoming much clearer that even generative AI, and the large language models that provide its engine, are imperfect, often inaccurate and prone to ‘hallucination’ – it can invent results.

If there is one way to ease professionals’ fears it is to make clear that the new technology requires knowledgeable professionals to understand and monitor the quality of data inputs for AI, and experts to verify the outputs. Much the same thing is required for any step change in the economy.

Chong anticipates a ‘reset’ from ‘overreliance’ on AI to one where people are recentred in the workflow process because outputs need correction. This will require high-level skills, and emphasis on the use of competencies in interpretation and judgment. He points out: ‘The hype says fear AI, but the reality is that humans will replace the humans.’