Thomas Braun - of the Executive Suite, explains how Workday's matching algorithm takes in candidates' personal data and resumes, feeding it into a model, where the scope for bias to exist seems inevitable? If it matches candidates to jobs using "cosine similarity." The problem emerges when an employer's own hiring pattern (e.g. repeatedly hiring people from certain demographics) teaches the algorithm to adjust its match scores accordingly. It will began surfacing only candidates from a particular region, state, or religious background, learning and amplifying bias from the employer's past choices. 

Watch Dr Cathy O'Neil, bestselling author on the risks of AI & pioneer of algorithmic auditing.  

In 2021, Cathy O'Neil audited HireVue's facial analysis tool, which assessed candidates in one-way interviews based on their facial expressions. That kind of tool is now 100% banned under EU law. The fact that damage can only be measured after it's happened makes companies who adopt it financially liable too. In the video, she calls all algorithms 'an opinion embedded in math,' and calls on us to inject ethics into the process of building algorithms. #WVH only injects your ethics, shares that with everyone involved, and will give you a Network Time Protocol stamped cryptographic hash of the codebase of your version of Quinn — protecting you from claims of systemic exclusion.