Human Resources – A Shift Towards artificial intelligence?

On 6 May 2025, the SRA authorised the first law firm providing legal services through artificial intelligence. Garfield.Law will provide an AI-powered tool which can assist businesses with the small claims court process, to aid in recovering unpaid debts. This demonstrates that specialised AI tools are increasingly seen as able to provide reliable guidance and, in certain limited circumstances, AI agents are trusted to carry out decision-making actions.  AI investors and proponents see widespread use of agentic AI (i.e. AI which can be programmed to make automated decisions) as a rapidly approaching future which will radically change every aspect of daily life.  There are clear potential uses of existing and hypothetical AI tools in human resources functions, but HR departments will need to be cautious in making use of AI, both to mitigate specific legal risks, and to not lose sight of the human in human resources.

AI tools and their application to HR

AI tools can be used for a variety of purposes for HR. Generative AI (such as ChatGPT) can be used to create content in response to the user’s input. HR may use this as a basis to help create job adverts and descriptions by submitting information relating to what you would like to be present in the output, and then tweak the produced result to ensure it is accurate and fits the business needs. AI can also be used to track performance trends of employees and provide a objective analysis of their performance to be used in employee reviews.  As these 2 examples demonstrate, AI’s use is wide ranging and can be used as a starting point to a larger task.

However, AI may also be used to perform simple tasks. Perhaps the most high-profile example an such a AI HR tool is the IBM Ask HR agent.  IBM Korea CTO Lee Ji-eun recently announced that AskHR has taken over 94% of simple HR tasks, such as approving holiday requests and payslips.  This has, according to CEO Arvind Krishna, led the company to significantly reduce the headcount of its HR teams worldwide and as a result, has led to further investment into other areas of the business.  They have emphasised that the roles of HR employees is evolving in order to cooperate with the use of AI systems. By implementing a AI system in which AI can deal with routine administrative tasks, it shifts employees’ focus into other projects such as ensuring they are satisfied with the policies in place within the business.

AI accuracy

Whilst the above may suggest that AI can take over roles or duties, AI tools have not yet solved the issue of ‘hallucinations’ where they produce inaccurate information with apparent authority and confidence.  This means that trained humans are still irreplaceable.  Decision makers need to be certain that both their “input” information and “outputs” are accurate.  With reference to the  tasks mentioned above, such tasks should be undertaken cautiously to ensure that language used in job adverts are not discriminatory and similarly there is no bias of metrics/factors used when undertaking performance trends. There are numerous examples where AI systems put in place by companies have provided unexpected results, placing the company in a sticky situation. However, AI will only evolve according to the information it collects from humans and if such information is discriminatory and/or inaccurate, this can have severe financial and reputational consequences for organisations.

For example, Air Canada’s chatbot incorrectly interpreted the airline’s policy and gave a customer the incorrect information stating they would be entitled to a potential refund – in a small claims court case, the court held that the company were bound by the chatbot’s response.  It shows that whilst there are existing AI systems which can perform these tasks, errors can occur and human interaction and oversight is still essential for the roles they are programmed to carry out, although AI use may still be beneficial as a time saver for your HR team.

Another example involves Estee Lauder, in which three make-up artists lost their jobs after being asked to take a video interview that was assessed by an algorithm. These make-up artists were told they were being made redundant partly on the basis of an automated judgment by a computer. Originally, they had been required to reapply for their positions. Such a case demonstrates the repercussions of unconscious bias in AI and where AI/algorithms going unchecked can significantly harm an organisation on both financial and reputational grounds.

The right to not be subject to automated decision making

In terms of legality and the risk of AI accidentally making the wrong decision, Section 49 of the Data Protection Act 2018 (“DPA 2018”) prevents data controllers from taking a decision which produces an adverse legal effect or otherwise significantly affects an individual, based solely on automated processing.

A not-insignificant number of employment disputes arise from the application of rigid policies to situations which do not neatly fit the contexts they were designed for.  These policies are necessary for consistent decision making, but can occasionally produce perverse outcomes if followed too rigidly.  The right in s. 49 of the DPA 2018 is, amongst other things, designed to allow a human to intervene and consider all available information and if necessary, diverge from a written policy and change the decision.  Automated tools, no matter how complex, are not yet able to do this.

While AI can be a useful tool in helping to assist with HR duties and responsibilities, particularly routine admin decisions, it is not yet capable of independent thought. It can make mistakes and its rigid use may cause more problems for your business than solutions. However, there are clear growing trends which suggest businesses and HR professionals in particular want to remain up to date and implement the use of AI, but it is important to remember that those decisions still require human interaction to ensure lawfulness and fairness to your fellow employees. With the use of AI in the HR sphere, the role of HR employees will evolve – the empathetic element of the role will remain but the more routine, admin elements of the role may be passed on to AI. We will continue to monitor developments in this area.

How we can help

If you would like to discuss the use of AI within your HR team, please feel free to contact our employment or data protection team. We can help advise you on how to ensure that the use of this technology maximises your efficiency while ensuring that the implementation is lawful. We also have policies available for the use of AI.