AI is being used increasingly by employers for recruitment, monitoring and workforce management. Whilst AI can improve decision making and boost workforce productivity, it can also bring potential issues and risks which need to be considered carefully.
Legal considerations
The traditional recruitment process has been evolving over recent years with AI now playing a key role in talent acquisition. Whilst AI can remove certain biases by focusing on the qualifications and skills relevant to the job description, it can potentially introduce other biases. AI algorithms learn through observing and repeating behaviours and so, if not audited regularly, can present a real risk of adopting or exacerbating unchecked biases. Specific types of AI, for example, facial recognition technology used at interview stage, can also present challenges. The technology has been found to provide less accurate results for female and ethnic minority candidates in some instances and result in them receiving lower scores. Examples such as this carry clear risks of discrimination complaints under the Equality Act 2010 which employers need to be alert to.
Monitoring employees through AI is another key concern. With hybrid and remote working being the norm for many employers, some are now relying on AI tools to monitor employees' activities, productivity and performance. Employees, however, have the right to privacy under Article 8 of the Human Rights Act 1998. Although it is possible for employers to justify any interference with employees' privacy rights, the legal scope of this justification is not always clear cut. In addition, employers must comply with the requirements of the Data Protection Act 2018 by ensuring any personal data obtained and processed concerning employees is done in a fair, lawful and transparent way.
Finally, there are further issues concerning the use of AI in workforce management processes and procedures which should be considered before and during use. For example, when using AI for the purposes of decision making concerning the dismissal of an employee, there is potential for unfair dismissal claims under the Employment Rights Act 1996 as well as risks under the other legislation mentioned above. In all cases, employers should use a human review to ensure any significant decision making is fair, reasonable and free from any errors or biases resulting from AI. Failing to do this may otherwise give rise to costly potential claims.
Practical considerations
A key way to overcome the challenges presented by use of AI in employment is to ensure there is robust human oversight. Ensuring any AI-led decisions are overseen by a human reviewer will allow any issues with AI technology or algorithms to be identified and rectified quickly. The implementation of specific AI policies governing what this oversight should look like in each workplace or decision making process can be helpful.
If monitoring employees, employers must establish monitoring policies which are reasonable and proportionate and do not unlawfully interfere with an employee's right to privacy. By adopting monitoring policies which establish clear practices and expectations, employers can ensure transparency in their communications with employees and minimise risk of breaching the relevant legislation.
AI systems used for recruitment purposes should be tested, challenged and updated regularly to ensure there are no issues which may cause biases. In addition, ensuring that systems and decision making is reviewed by humans within relevant teams will minimise the risk of developing unchecked issues.
Summary
Although AI may save employers time and money in some areas by accelerating recruitment, streamlining decision-making processes and efficiently monitoring employees, organisations must ensure they are retaining an appropriate level of human oversight to reduce risk. Complying with relevant legislation should remain a priority for employers, not only to avoid legal risk but to also ensure trust within the workforce and to maintain positive employer brand and reputation.
