These are key risks AI poses to employers, says legal expert

Firms must really get to grips with the potentially serious risks posed by the use of artificial intelligence (AI), according to an expert in employment law.

Euan Smith, senior office partner at law firm Eversheds Sutherland in Edinburgh, said the topic has been coming up in conversation with clients, “because it's such a buzzword at the moment” – and the organisation is working to ensure they have in place suitable guardrails.

Eversheds Sutherland noted, in a global report on the technology last month, that in the UK, the Artificial Intelligence (Regulation) Bill has passed through the House of Lords and has now been sent to the Commons, while the Information Commissioner's Office has launched the third chapter of its consultation series on generative AI – tech that can produce, say, text and images.

Smith cited, for example, risks related to employees using AI to ask natural language questions and getting answers that they present as gospel in official documents presented to colleagues and/or clients. “One of the problems with that is that [such technology is] not hugely accurate yet, it tends to just be an aggregation of sources from websites. And therefore, there is a need to verify all of the information that we get through use of any of those kind of AI engines.

“There's a bit of work right now in terms of educating employers about this, putting in place policies, doing training for employees about how to use these tools safely, in a way that's not going to expose their organisation to any greater risk.”

He also highlighted potential dangers in businesses using AI tools for recruitment, for example to screen reams of CVs. While that can evidently save much staff resource, Smith said AI models have to be taught with a data set that may include an in-built bias that is then amplified by the technology. He explained that a firm might be looking to hire into a department that is currently mainly staffed by white males under 60, which could mean the AI would then only select CVs meeting this brief. This could then exclude “whole groups of people for no good reason… and that gives rise, quite obviously, to the potential of legal claims”.

The employment lawyer added that there is also a risk of claims that arise under data protection rules that mean an employee must not be the subject of automated decision-making, and a breach can lead to a fine that is a percentage of global turnover.

AI is becoming increasingly mainstream, 'but there are inherent risks with it, which employers need to be alive to,' says Smith. Picture: Lisa Ferguson.

Eversheds Sutherland also noted in its report last month that the UK Government had published new guidance on procuring and deploying AI technology and systems responsibly in HR and recruitment, warning that such technologies can expose companies to novel ethical risks.

Smith said there is “clearly a burst of enthusiasm” for AI right now, having rapidly gone from the preserve of tech specialists to widespread availability, and with obvious benefits for businesses. “[AI] is becoming more predominant – it’s growing all the time in terms of its extent, and how many people are using it and the use cases that they are applying it to. But there are inherent risks with it, which employers need to be alive to.”