By Justin Birch, Executive Search Director, New Street Consulting Group
“In April 2023, employers in New York will be required to conduct an annual AI bias audit of automated employment decision tools and publicly disclose related AI hiring metrics. These tools use various criteria to assess candidates from algorithms built to find the best applicants to software that evaluates body language.
“In addition to requiring AI tool audits, the law states employers must make the results publicly available on their websites. Employers will now also have to consult technical experts, as well as attorneys who can spot the potential for discrimination complaints when chosen software is used.
“Organisations that don’t meet these legal requirements will be subject to a $500 fine for the first violation and $1,500 for each subsequent violation. The reasons for this decision are to create more transparency in how AI is used in hiring and reveal any apparent disparate impact.
“Will these new laws start to make their way over to the UK and are we considering the repercussions that might occur from using such software regularly?
“While no official legislation has yet been passed, it seems many of us in the talent acquisition industry are in two minds about the use of AI and its effectiveness.
“AI should make hiring more efficient by filtering from a larger candidate pool. However, it can also miss or exclude good candidates if the rules and training data are incomplete or inaccurate. For example, AI may be able to support certain areas of recruitment, but it will struggle with the likes of culture fit which is becoming increasingly important over hard skills, especially in senior leadership roles.
“I find human interaction to be irreplaceable when recruiting for a role where there is no set job specification. This often occurs at senior level where an organisation realises there’s a gap that needs filling; perhaps a problem has been identified, such as a need for transformation, the implementation of new technology or board advisory.
“In these cases, a job description doesn’t really exist, and part of the interview process is to work with top class candidates to explain the issue, allow them to offer examples of their experience and solutions and by doing this flesh out the role and responsibilities. It is unlikely that AI can do this as there is very little data to start with. It is also important to have some sort of balance and the ability/freedom to choose somebody who might, in some cases, be considered a non-traditional candidate fit.
“Finally, I really do believe human interaction is necessary to fill leadership roles. The more senior the role, the more vital it is to the company and the more diligent the recruitment process.
“Companies, roles and candidates have many nuances that aren’t documented on a CV, job description or GlassDoor. Take motivations, for example. A candidate may say they are motivated by X, but an assessment may point to them being motivated by Y. Only when you layer this in the context of all your other interactions with the candidate can you begin to understand what truly motivates them and whether that’s appropriate for the role.
“AI can’t replicate this, nor can it truly understand the subtleties of direct human communication.”