Artificial Intelligence in the Recruitment Process – Risks & Recommendations

Artificial Intelligence (AI) is a huge topic at the moment across almost all industries, not least in relation to its effect on recruitment and selection within business. Regardless of your view on AI, it is clear that it is here to stay. However, there is a need to consider the potential implications of using AI, particularly in areas where AI has the potential to make or inform decisions about individuals, such as those made during the recruitment process.

Employers are increasingly using AI tools during the recruitment process to:

• Perform reviews of CVs and application forms;
• Search prospective employees’ social media for key phrases or terms;
• Analyse tone of voice or facial movements during interviews; and
• Perform automatic filtering of candidates through online assessments and tests.

At this early stage in the development of AI, employers should err on the side of caution as AI tools lack human intuition, which could cause issues when an employer is trying to justify their selection decisions. If an employer decides to use AI software, it must have a clear understanding of the algorithm they are using and how it works. There are obvious benefits to using AI, not least its capacity to efficiently review a large volume of applications and its apparent ability to make unbiased decisions during the recruitment process.  However, it also comes with significant risk.

Learning from Past Mistakes

The main legal issue in using AI during the recruitment process is the potential for possible unlawful discrimination arising out of unfair decision-making. Recently Uber was criticised after its AI facial identification software reportedly failed to accurately identify dark-skinned faces, resulting in many workers being unable to access its app and apply for jobs. Further, Amazon built its own automated CV screening algorithm and trained the algorithm using its recruitment data, collated over ten years. However, given the nature of AI, it analysed Amazon’s previous recruitment decisions and taught itself that male candidates were preferable to female candidates. Due to this, it reportedly penalised CVs that included the words “women’s” and downgraded graduates of certain all-female colleges. Amazon later abandoned their automated CV screening project in 2017. This project demonstrates how algorithms can perpetuate human biases when they are trained to learn from historical approaches to recruitment.

AI & Discrimination in the Recruitment Process

A recent TUC report on the legal implications of using AI found that AI can result in discrimination claims. An algorithm can be seen as a “provision, criterion or practice” within the meaning of equality legislation in place to protect individuals with protected characteristics, i.e. disability, age, sex, etc, for the purpose of bringing an indirect discrimination claim. Indirect discrimination is when a provision, criterion or practice (PCP) has been applied by the employer which puts a particular employee with a protected characteristic at a disadvantage compared to other employees who do not have that protected characteristic. An employer has to prove that the PCP was a proportionate means of achieving a legitimate aim in order to justify its application of AI.

Recommendations for using AI in the Recruitment Process

Below are five recommendations employers should consider if using AI during the recruitment process:

• It should not be used to replace human line management, and managers should always have the final responsibility for any workplace decision;
• Line managers should receive training on how to understand the algorithms;
• Algorithms should never be used to mask intentional discrimination;
• Employers should be transparent with employees (and prospective employees) about when and how algorithms are being used and how they can be challenged; and
• Employers should audit what is being measured before implementing the program, and on an ongoing basis.

Finally, employers will also need to consider the GDPR implications of using AI in recruitment and provide appropriately drafted privacy notices to candidates to comply with Data Protection obligations.

Should you require any information or advice in relation to the matters raised in this article, please do not hesitate to contact our Employment team.

This article has been produced for general information purposes and further advice should be sought from a professional advisor.