Much (and I mean much), has been made in recent months of the growing involvement of generative AI in recruitment.
It’s understandable.
We can’t ignore that it’s here, because it rise has been nothing short of stunning.
We also can’t also ignore the fact that it will continue to grow rapidly in the next few years, with the global HR GenAI market size expected to be worth $1.67 billion by 2032, representing a CAGR of 15.4% over the coming years.
But could this emerging technology – one that often has much suspicion around it – actually pave the way for a much fairer ecosystem for candidates and businesses alike, and spell the end to the age-old problem of unconscious bias in recruitment?
I believe generative AI does have the potential to overcome age-old challenges that have hindered the efficiency of recruiters.
While there’s little doubt GenAI will pave the way for significant time-saving processes within the industry, it’s the potential for removing unconscious bias that’s exciting, and which could transform HR as we know it.
To me, what more relevant are the new questions that this poses: like whether this new faction of artificial intelligence deliver a more objective recruitment process? Or could the technology throw up fresh challenges for recruiters?
So what are the answers?
Let’s first take a deeper look into the transformative potential of embracing GenAI in recruitment:
Recruitment has long accepted human error as an industry mainstay, with factors like tiredness, data entry oversight, and both conscious and unconscious bias seeping into operations.
According to BrightTALK data, 79% of HR professionals believe that unconscious bias exists in recruitment and succession planning, and overcoming this flaw can be particularly challenging.
This is where generative AI enters the fray. The ability of artificial intelligence to actively counter unconscious bias could be a revolutionary quality for HR and finally put an end to unfair recruitment practices.
Generative tools can actively remove personal identifiers like candidate names, education, and other information that could see unconscious bias forming surrounding an applicant’s gender, race, age, interests, or background.
This empowers recruiters to judge applicants purely based on their suitability and personality fit objectively.
Recruitment today hinges on interviews. However, during the interview process, candidates produce significant volumes of unstructured data.
This conversational level of data is open to interpretation by the interviewer and thus subject to unconscious bias.
Whether an interview is being conducted on the phone or in person, many different variables can lead to recruiters missing out on valuable insights into a candidate’s suitability.
For instance, a valuable piece of information may be missed within a larger point being made by the candidate, or an interviewer could negatively misinterpret another valid point for several reasons.
Artificial intelligence can remove these unstructured barriers and create transcripts where important information is extracted and summarized autonomously.
Crucially, this will save recruiters hours and offer a brand-new way for firms to gain valuable insights from their interview process.
We’re also seeing use-cases emerge for AI to host candidate interviews on behalf of recruiters, which can be a useful tool for job roles experiencing high volumes of applicants.
Here, generative AI can summarize a candidate’s suitability by setting questions based on its interpretation of their CV and analyzing their responses.
Again, unstructured data like facial expressions, body language, and tone of voice can be assessed for a more comprehensive understanding of a candidate’s suitability.
Generative AI can flourish with natural language processing (NLP), and for recruiters, this means an unprecedented level of analysis of job descriptions and next-generation candidate screening processes.
In a similar way to the extraction of unstructured data, NLPs can scan job descriptions and automatically extract candidate skills, experience, and qualifications to determine their suitability for a role.
Although these mechanisms exist to an extent among HR tools, generative AI can understand the context and semantics surrounding job descriptions and accurately highlight the best candidates for a role in a more effective way.
Additionally, this natural understanding of prompts can play a significant role when recruiters want to generate a job description.
Unconscious bias can seep into many areas of recruitment, and this means that using AI to generate job descriptions can be an excellent way of advertising a role’s requirements to every talented candidate, as opposed to the ones that the recruiter is imagining when crafting their description.
At this stage, it’s important to highlight that generative AI and machine learning techniques in recruitment will only be as objective as their programming or source materials allow.
With this in mind, GenAI could replace unconscious bias with emergent bias, which means that ML algorithms could learn the biases that recruiters commonly experience from source data.
This could amplify unconscious bias and make it harder for recruiters to spot throughout these processes.
Any business or agency adding generative AI to their HR operations will need to stay on the lookout for emergent biases and ensure that they monitor the material used to train the ML algorithms used to power generative AI solutions within the industry.
Although generative AI is by no means flawless in tackling bias in recruitment, the right level of training and programming can turn the technology into the perfect objective assistant for recruiters and leverage many time-saving tools to build an industry that’s fairer for all participants.
Generative AI will be the most disruptive technology of the decade, and for the HR departments that unlock its potential, a more efficient and accurate recruitment process awaits.
