How AI is detecting unconscious bias in recruitment

Vivek Doraiswamy, Co-Founder of MeVitae, explains how artificial intelligence can strip away prejudices and bias to ensure equality and diversity in recruitment.

To succeed in a fast-paced, hyper-competitive global marketplace, companies must innovate, but a technology skills shortage is looming. As a result, 94% of the fastest growing tech firms state that finding, recruiting and retaining tech talent is a growing concern.

Companies need a whole new and diverse talent pool (with respect to variables including ethnicity, gender and disabilities) for economical survival and growth. Companies declared a ‘commitment to equal opportunities’ in the 1990s, yet diversity levels haven’t changed since the early 2000s. Twitter’s tech staff is only 2% Hispanic, 2% black and 28% women, whilst Microsoft show 5%, 3% and 24%, respectively. Unconscious bias is the leading cause for the lack of diversity within the workforce.

What is unconscious bias?

Google was the first to call out unconscious bias for contributing to the systematic lack of diversity in the technology industry. Sheryl Sandberg, Facebook’s COO, states ‘The most important thing is to correct for unconscious bias to help us build stronger, more diverse and inclusive companies’.

Unconscious biases are the automatic, mental shortcuts used to process information and make decisions quickly, growing people to social norms and stereotypes. Unconscious bias can prevent individuals from making the most objective decisions. They can cause people to overlook great ideas, undermine individual potential and create a less than ideal work experience for their colleagues.

By understanding unconscious bias and overcoming it at critical moments, individuals can make better decisions – from finding the best talent (no matter what their background) to acknowledging a great idea (no matter who it came from) and build a workforce that supports and encourages diverse perspectives and contributions. Harvard University has developed the Implicit Association Test (IAT) to spot unconscious bias and this can lead to bias training. However, strong evidence from Intel suggests diversity training turns out to be an ineffective approach to improving diversity and eliminating managerial bias.

How does this affect the recruitment industry?

According to BeHiring, companies receive applications within 200 seconds after a job is posted, and an average of 250 CVs are received for each job position. This demand requires recruiters to read through CVs quickly as competition for talent increases.

In 2012, TheLadders conducted the first ever formal, quantitative study of recruiters’ on-the-job behaviour, where eye-tracking was used for 10 weeks, concluding that recruiters only spend a mere six seconds reviewing a candidate’s CV. In 2015, research shows that just 8.8 seconds is spent studying any one person’s CV in a process that has become ‘Tinderised’. What do they look at during that time and how do they make choices? How do you detect and correct for unconscious bias?

What’s the alternative?

This is where MeVitae comes in, building the first-ever technological alternative to ineffective unconscious bias training, working with the Oxford University Physics Department, where I lead and supervise students including Ilya Lapan and Xiaoli Huang. The vision is to end diversity as a buzzword, be a force for breaking glass ceilings and close the technology skills gap to accelerate innovation.

Mevitae is working with Oxford University to tackle unconscious bias

Mevitae is working with Oxford University to tackle unconscious bias

Eye-tracking devices, neueoimaging and Emotiv Electroencephalogram (EEG) coupled together, will allow MeVitae to measure how employers’ brains react to Harvard University’s IAT test, different CVs and sentences. The neural behaviours of hiring employers will be decoded to determine patterns on what employers look at in CVs and build a bias correction algorithm that scores, ranks and shortlists candidates for jobs, irrespective of social class, race, ethnicity, gender, disabilities, etc. through big data and artificial intelligence (augmented human intelligence). This will highlight the unconscious bias hindering diversity in the workplace.

This innovative solution, due to be released this spring, has already garnering interest by top innovative and technology companies both in the UK and Silicon Valley.

Riham Satti, CEO of MeVitae, states that “this is the first step towards disrupting the human capital market and solving the biggest recruitment challenges, with the vision of unleashing human potential through cognition.”

Vivek Doraiswamy is Co-Founder and CTO at MeVitae. You can contact him at If you would like to hear more, follow them on Twitter @mevitae. Don’t forget to follow us too @DigiCatapult.

Get involved with our AI events and activities, including the IBM Watson AI XPRIZE – a $5m competition to solve the world’s biggest challenges with artificial intelligence. 

Leave a comment