It is increasingly clear that protections under existing employment law may be insufficient to deal with the challenges posed by the increasing use of artificial intelligence (AI) and Automated Data Management (ADM) in HR processes.
For instance, many companies now use AI providers such as HireVue to carry out recruitment tasks previously done by humans, including advertising, background checks, “skimming” CVs for key words and even first level interviews. However, the use of such technology carries the risk of discrimination that may go unnoticed if there is no human presence to oversee it. Critics of HireVue have demonstrated that its face recognition software (now removed) discriminated against disabled candidates, including those who were partially sighted, on the basis of their facial responses.
While there is protection from discrimination under existing law, it may be difficult or impossible for candidates to prove that has taken place if they are not aware that AI played a part in the decision-making process.
This issue extends to the use of AI in, for instance, monitoring employees, or selecting them for redundancy. Without the oversight of a human manager, incorrect and unfair decisions can result. The commonly held assumption that AI and ADM decisions are impartial and unbiased is simply not true. Humans are responsible for providing the original data and, whether intentionally or unintentionally, that can lead to skewed or discriminatory decisions later on – the “rubbish in, rubbish out” principle.
Inappropriate use of AI/ADM strikes to the heart of the employee/employer relationship, which is founded on the principle of mutual “trust and confidence.” Under common law it is understood that the employer, as the more powerful party in the contract, has an obligation to consult and explain a decision-making process.
However, there is currently no right to demand that a final decision is made by a human rather than AI, nor to ask for proof that a human has had oversight over the process. If employers are not transparent and open about how AI/ADM is used, at what stage, and the role that humans take in the process, then it will be increasingly difficult for employees to prove unfairness or discrimination.
The issue is even more pressing for those who do not have the protection of employment legislation, such as the Just Eat self-employed couriers who recently found themselves thrown off their work app by AI for overpayments as small as £1.44. https://www.theguardian.com/business/2023/apr/22/fired-by-ai-just-eat-uk-couriers-deactivated-for-minor-overpayments The company has claimed without evidence that the decisions were overviewed by a human manager, and the couriers have no right of appeal.
Some legal protection exists regarding decisions informed by inaccurate data, as under the UKGDPR all data processed as part of AI/ADM must be accurate (Article 5(1)(d)). However, complaints under the UKGDPR cannot be pursued in the Employment Tribunal; the only punishment for breach of UKGDPR is a fine for the company itself, which is of little comfort to an employee or worker subjected to unfair treatment.
So how do employers and employees protect themselves from the potential problems of “inhuman resources”?
Businesses must ensure that they are open and transparent about the use of AI in any context. It should not be used for high level decisions such as dismissal or redundancy, unless those decisions can be sensibly explained and understood by employees, and where AI is being used in the interview process, that should be made clear to applicants and some form of appeal offered where discrimination appears to have occurred.
Complex technological solutions can have a dangerous mystique, and even the most tech-savvy managers are unlikely to fully understand the capabilities of their AI systems or anticipate possible pitfalls. The revelations of the Post Office scandal, where the Horizon computer system was held to be more trustworthy than the humans tasked with using it, is proof of how damaging complacency can be.
Prevention is always better than cure and now is the time for the Government to look at how employment law can be adapted, before AI and ADM become an unavoidable part of the employment process.
Our solicitors are always on hand to give you friendly, professional advice. Please contact Zoe on 0203 858 9765 or email zoe@mulberryssolicitors.com. Mulberry’s has offices in Brighton and London.
This article is for general information only and does not constitute legal or professional advice. Please note that the law may have changed since this article was published.
Leave a comment