AI and employment law: What HR leaders need to know

As artificial intelligence takes on more HR tasks and is integrated into workplaces, employers must balance potential productivity gains with new risks and compliance issues.

Cilla Robinson, a partner in the employee relations and safety team at King & Wood Mallesons, spoke to HRD about some of the key employment law issues for HR leaders as work and decision-making shifts to AI.

AI and recruitment 

Recruitment is one of the HR functions where AI has been rapidly implemented, with a promise of removing conscious and unconscious bias from the hiring process.

But Robinson warns that “while on paper” it seems like a way to remove the bias that humans can bring, AI can also replicate or amplify bias.

“AI is a tool that learns from the information it has at hand. If you get any organisation and look at the last 20 years of successful candidates, that might skew them towards men, at least in certain industries,” she said.

“The key thing for HR professionals to remember in the context of AI, is if you put in data that's biased, it's going to exacerbate bias, without careful prompts to avoid this outcome”.

“Obviously, the upside is that there can be a huge time saving in terms of automation of screening and sifting through applications, so I'm not trying to scare off HR professionals. It just needs to be used with care to ensure that you're getting the best out of AI in that context.”

Robinson says AI tools need to be audited for bias to ensure an employer is not engaging inadvertently in discriminatory practices.

AI in terminations

Robinson said it is a “very dangerous game” to rely on AI for decisions when it comes to termination decisions and HR leaders need to ensure AI is used in an appropriate and ethical manner.

She said these decisions often required nuance, and a balance of competing factors, and AI was not always going to perform well in this area.

“There might be a valid reason to terminate, but nevertheless it's unfair because this person has had 25 years of perfect conduct or an unblemished performance record. Those types of considerations that a human will have regard to are sometimes just missing if you're asking AI to do that sort of discretionary work and decision making, without regard to the full background set.”

Gaps in the law

Robinson points out the current discrimination law talks about ‘a person’ discriminating, making it unlikely that a court could find employers liable if they used algorithms that made discriminatory decisions.

“AI actually falls between the cracks in terms of where the law is at, so the law needs to catch up to avoid potentially disastrous DEI outcomes where no one is liable,” she said.

There is also no requirement at law presently to disclose how AI is being used in recruitment.

“That does make it quite difficult, as a candidate, to work out if AI is discriminating against you,” she said.

“It's just something to be aware of: if the law changes and we see obligations to disclose AI usage or the introduction of agent liability, like we see in emerging case law in the US, then HR might need to start advising candidates where AI is being used in recruitment and being very careful to ensure discriminatory or other perverse outcomes don’t arise.”

AI and privacy

Robinson says it’s important for HR leaders to ensure their use of AI complies with privacy laws.

“AI systems often require large data sets, and sometimes those data sets can include sensitive employee information,” she said.

“There is theoretically a situation where mismanagement of data could lead to significant data breaches of personal information for example and reputational as well as legal consequences that follow as a result of that.”

How AI is used in the workplace

Robinson believes the most challenging issue for HR is managing the use of AI in the workplace by employees.

“That can quite often become a conduct or performance management issue where the supervisor doesn't realise that their employee is giving them work that is actually being created by AI,” she said.

AI can also “hallucinate” and create work that is completely fabricated, presented in a confident and factual manner.

“I think that's going to be a huge area in terms of the workload of HR to try and manage that practical side of AI and the impact in employee relations. Then you have the overlay that many unions are aggressively seeking to limit the use of AI and mandate consultation before any AI is implemented within businesses. This has both practical implications on day-to-day operations as well as delaying potential productivity gains that AI can deliver,” she said.

“You need to understand who is using AI and when, and have the basics in place such as an AI use policy.

“I think it's really critical that the company and the HR staff are clear on what the expectations are, communicate those expectations to the employees in the workforce and then ensure that there is consequence management for people who use it in an unauthorised way.”

“HR professionals need to understand the limitations of AI so they can then ensure there are checks and balances in place to ensure the accuracy of any information.”

Surveillance and monitoring of employees’ output

Robinson says monitoring performance can create concerns about a “Big Brother element” to oversight.

“The role of HR and employment lawyers is to ensure there's a balance between the need for oversight and the benefits that AI can bring to monitoring performance and productivity, but also to respect employee privacy. Surveillance has to comply with privacy and workplace surveillance laws and be ethical,” she said.

“Employees need to know they are being monitored, and from a surveillance perspective they need to consent to that use.

“Practically from an HR perspective, there has been an increase in stress claims off the back of that kind of monitoring. I think acknowledging that such monitoring, whilst good for business, can have an adverse impact on the humans that are doing the work. Management of this risk to wellbeing is also an important HR consideration.”

Psychosocial hazards: AI's impact on workforce dynamics

If AI leads to workplace changes such as workforce reductions and restructures or even just changes to the way work is performed, employers need to consider the psychosocial risks to their workers. Decreased staff morale and increased stress claims are not uncommon.

Robinson said it was important HR was at the table during the implementation phase of AI-related changes, particularly if it is linked to cost and workforce reductions, to ensure the “human element” was being considered.

“I think just understanding there are challenges so that you can emphasise the risks, the opportunities, the importance of having guidelines and to help the organisation stay across the legal landscape as it continues to evolve, because you know the law is always catching up to the practices when it comes to technological advances.”