We are aware that bias in the workplace is a major issue. Around 30% of UK employees have experienced or witnessed bias in their career. Gender bias is the most common, followed closely by ageism, and racial discrimination.
These issues persist despite years of effort to promote equality and diversification, especially in professional service industries such as finance and law.
Artificial Intelligence is increasingly being used by organisations to streamline their recruitment processes and address inefficiencies. As someone who has worked in recruitment for over 25 years and educates leaders about inclusive hiring practices, while AI promises speed and objectivity I am worried that it poses a risk of perpetuating or exacerbating biases embedded in the workplace culture.
AI Models: Pitfalls
AI systems that are used for recruitment are usually trained using historical data from previous hiring practices. AI systems that are trained with biased data or algorithms will continue to perpetuate biases in hiring. It is especially concerning in industries like finance and law, where the lack diversity in senior management reflects years of unequal opportunities. A major tech ecommerce company was the victim of AI bias.
The company created an AI-driven tool for hiring that was unintentionally biased against women. The system, which was trained using 10 years’ worth of resumes and a large number of applicants, penalised those applications that contained words such as “women’s chess clubs” because the majority of previous employees were men. Researchers at Carnegie Mellon University found that Google’s AI platform showed more job ads to men, than women. This reinforced existing income disparities.
These incidents, while rectified quickly by the organisations involved, serve as warnings about what can happen if human oversight is removed during the recruitment process.
Human Factor
It is important to maintain human involvement in the recruitment process, especially at the beginning. These examples highlight the importance of real people being involved in recruitment right from the start. You will be at a disadvantage if you only rely on AI for your recruitment strategies. Your competitors may use other methods to find talent, such as networking. AI is good at quickly sifting vast amounts of data and identifying applicants who meet certain qualifications.
It struggles to take into account the context and nuance that humans bring when making decisions. An AI system may favour candidates who attended a certain university, based on the performance of previous hires. However, this could exclude individuals with less traditional backgrounds. AI is inherently unprepared to assess the potential of candidates that don’t conform to a traditional mold.
An algorithm that is trained to favor linear, continuous career progression may overlook someone who has taken a career break in order to raise a child or changed careers.
Minimising Bias
A diverse recruitment strategy will help you achieve your goals. Companies must have a diverse and comprehensive recruitment strategy because of the risks associated with AI bias. It is important that organisations take deliberate measures to ensure that all candidates are given equal chances to succeed. This is about more than filling a quota. It’s important to recognise the value of different perspectives and experience in an organization.
Auditing recruitment practices and identifying bias patterns is crucial. These data are important for the design of future AI systems that are fairer and can also be used to guide manual recruitment processes. In order to do this, companies need to train AI using diverse data sets. This will expose AI systems to a variety of candidate profiles and reduce the risk of bias in decision-making.
Let’s get back to the human recruitment. To ensure fairness and to combat unconscious bias, diverse panels should be involved in the hiring process. Prioritizing human oversight is important. AI should not replace human judgment, but rather complement it. I believe that recruiters need to review AI recommendations and make final hiring decisions after a thorough evaluation of every candidate.
Use AI Responsibly
AI has the ability to improve recruitment efficiency and provide insight. We must be responsible with AI. This means that organisations must recognise the limitations of AI, and invest in systems and strategies which prioritise inclusivity and fairness. We must remember that when we discuss strategies, human-led ones, like the new wave of network we are leading at DTN, can help to break down unconscious and organisational biases and put the right people before business leaders and key decision makers. This is not the traditional networking we are used to!
The goal of AI in the recruitment process should not be to replace humans, but rather to improve human decision making. Combining the speed and accuracy of AI with empathy and insight from human recruiters can help organisations create truly equal workplaces, where bias is not tolerated and everyone has a chance to flourish. Companies cannot afford to be held back by bias in a world where talent has become the most important asset.
