Using AI in recruitment: a guide for employers

Tom Clarke, Associate, Employment Law & HR

Workplaces are slowly but surely implementing AI tools to help with day-to-day tasks and a recent government announcement has made clear that the use of AI across a variety of sectors is set to dramatically increase. Changes to employment rights are also likely to see increased pressure on effective recruitment. This article serves as a brief guide on using AI in recruitment, and as well as the potential risks to be aware of.

How to use AI in recruitment

1. CV scanning: At the initial application stage, employers may receive hundreds of CVs which are difficult to sift through quickly and effectively to ensure the best candidates are considered. AI tools can automate this initial recruitment stage to search for and prioritise CVs containing key words relevant to what the employer is looking for, for example, determining whether the candidate has the minimum qualifications required. Using this tool may help employers to minimise not only the time spent manually screening each candidate but also any human bias, provided it is used correctly.

2. Scheduling interviews / AI monitored video interviews: When it comes to interviewing candidates, many employers will choose to interview either face to face or via a video call. AI diary tools can help to schedule these interviews around existing schedules and in accordance with candidate availability.
   Alternatively, some employers now choose to implement video interviews performed or assisted by AI. Using some AI software, the candidate records their response to a question within a time limit and AI can analyse and score this response.
   Other forms of AI software can analyse tone of voice, language and facial expressions during live video interviews, compare the traits it identifies to specified job requirements, and then provide scores for each candidate. Although these can be useful tools, it is always advised that AI is not solely relied upon to score candidates in order to avoid bias.

3. Analysing candidate performance: Another potential use for AI is to rank candidates depending on how they have performed in certain set tests, or more generally at each stage in the recruitment process. This is particularly helpful at the final stages of the recruitment process to give a more objective overview of each candidate, reducing the risk of human bias.

4. Employee onboarding: AI also has potential use in the employee onboarding process. AI can direct new employees to appropriate resources if they need help finding their employer’s policies for example, and can keep a record of any training the employee has completed.

What to be aware of when using AI in recruitment

1. Equality and diversity: The input into AI software will affect its output; this means that if there is bias already prevalent in the AI input, it will inevitably be reflected in the output.

As an example, in 2018 Amazon used AI in their recruitment algorithm to prioritise applicants who would be a ‘good fit’ for the company based on their top performer profiles. These profiles were predominantly men at the time, which led the AI algorithm to learn that male candidates were preferable to female candidates for Amazon, and therefore it started to penalise female CVs.

The Information Commissioners Office guidance on using AI in recruitment recommends that the AI tools are regularly monitored so that action can be taken swiftly to address any fairness, accuracy or bias issues. Although recruiters may not always notice any bias in their AI input, it can be significantly amplified in the output, as the Amazon example shows. The key is minimising this, acting on any discriminatory outputs swiftly through regular monitoring, and keeping relevant policies up to date to reflect the employer’s continuing commitment to equality and diversity.

2. Data protection: The processing of candidates’ personal data is also a potential risk. Employers must have a reason to lawfully process a candidate’s personal data. With the addition of AI, employers must ensure full transparency with candidates as to how their information will be processed by AI and should be mindful of not inputting an individual’s personal data into software where it is not protected.

The ICO requires recruiters to inform candidates about how their information will be processed and carry out comprehensive risk assessments. It also recommends that recruiters collect only the minimum possible information to achieve AI’s purpose.

Summary
Employers are continuously adapting to a new technological landscape, and although there are risks associated with the use of AI in areas such as recruitment, it can certainly prove to be a useful tool if operated carefully and with regular monitoring.

For more information and tailored advice on using AI, contact Tom Clarke.