Can AI Prevent Unconscious Human Bias in the Recruitment Process?
Using Artificial Intelligence (AI) in recruitment should, in theory, prevent human bias in the
recruitment process. Human bias is a major problem in the workplace and a barrier to diversity
and inclusion. Internal training and awareness of this issue can help but humans have inherent
unconscious biases that can be difficult to overcome. Unconscious bias is defined as “learned
stereotypes that are automatic, unintentional, deeply ingrained, universal and able to influence
So how can we solve this very human problem? Turning to technology and artificial intelligence
(AI) can solve this and many unconscious bias training programs have been created with this
goal in mind.
Facebook designed a webpage to make unconscious bias training videos widely available and
Google has put about 60,000 employees through a 90-minute unconscious bias training program.
Last year, 67 percent of hiring managers and recruiters surveyed by LinkedIn said AI was saving
them time. So how can it be adapted for the recruitment process?
How Can AI Be Used in the Recruitment Process?
By applying AI techniques such as machine learning, sentiment analysis and natural language
processing you can automate and streamline parts of the recruiting process. Areas where
unconscious bias has impacted hiring such as candidate sourcing, screening and outreach can
then be improved.
Applying AI in Job Advertising
Job adverts have to be posted according to current legislation, you cannot usually refer to ‘protected characteristics‘ (such as a person’s sex or race) in a job advert. Examples that could also be discriminating unfairly against protected characteristics include:
- – Terms like ‘recent graduate’ or ‘highly experienced’ – these can discriminate against age
- – Terms like ‘barmaid’ or ‘handyman’ – these can discriminate against someone’s sex
- – Language requirements – ‘French sales rep’ could discriminate against race, so ‘French-speaking sales rep’ would be more appropriate
- – Clothing requirements could discriminate against someone who for religious reasons covers their head or face
Employers and agencies may not realise their job adverts can also put off potential applicants with certain words. Artificial Intelligence can ‘de-bias’ a job posting by using sentiment analysis to identify exclusionary language and offer alternatives to attract a more diverse talent pool.
CV Screening – Using AI to Overcome Unconscious Bias
AI can screen CVs instantly and by using machine learning, overcome bias by knowing the qualifications needed for a role. This looks at CV data and not at certain areas such as where someone is from or which university they went to. AI can be programmed to ignore demographic information such as gender, race or age or socioeconomic information (address or school).
So recruitment processes can be analysed for bias by looking at demographic breakdowns of hiring and comparing it to the total candidate applications. AI can find out where bias is happening more quickly than a human but we need to find ways new technologies can offer real opportunities to improve HR and recruitment functions. While technology can address the issues at stake, individuals still have to cope with solutions to address any biases that may affect diversity in the workplace.
NICOLA MONGON is the MD of Spotlight Recruitment, the leading marketing recruitment agency in the UK since 2008. Read our 5* Google reviews from candidates and clients to find our more. If you would like to discuss recruitment, you can reach Spotlight on 020 3008 4254 or firstname.lastname@example.org