When it comes to artificial intelligence (AI); its implementation, intended usage and outcome are heavily discussed, analysed and often critiqued, says Colin Willis.

The AI debate is constant and for good reason. New forms of generative AI, such as Chat GPT and Dall-E 2 have accelerated technological advancements and been implemented across a wide array of industries, reigniting the need for vigorous testing and regulation to ensure all forms of AI remain ethical.

This is especially true when considering its implementation in hiring. However, within this debate, there has been some confusion over the difference between generative AI and Natural Language Processing models.

As such, within this meaningful dialogue surrounding hiring and AI, emerges an echo chamber of misconception and conflation that has arisen primarily due to a limited understanding of the research that underpins regulated forms of AI. However, common misunderstandings about the implications of ethical AI often lead to sweeping binary generalisations about the pros and cons of humans vs. machines in AI.

Instead, people can breathe a sigh of relief, realising that misconceptions fail to acknowledge the myriad benefits that arise from the harmonious fusion of the two. 

Misconception: AI is inherently biased

A commonly held misconception surrounding AI in hiring is that it is inherently biased. Accessibility and inclusion in the hiring process are crucial factors, therefore, mitigating the opportunities for bias and exclusion should be central to assessing all recruitment strategies, whether or not AI is being implemented. The concern regarding AI in hiring is its potential to perpetuate existing biases within businesses. For example, consider an organisation with a historical bias of exclusively hiring men with degrees. The worry is that the AI algorithm might be trained on these biased “models of success”, leading to a reinforcement of such biases. Additionally, there is a fear that an irresponsible programmer could manipulate the algorithm to achieve specific outcomes, serving their own agenda.

However, this is not reflective of reality as AI hiring algorithms are developed with a combination of engineering and I/O psychology principles. The involvement of IO psychologists in the algorithm’s development is crucial. Their scientific expertise and ethical training enable them to rigorously evaluate the data before finalising the algorithm. Furthermore, IO psychologists hold the responsibility of assessing the algorithm’s output before its implementation within the organisation. They also continue to monitor its effectiveness and fairness after it has been put into use. The combination of the

engineers and I/O psychologists ensure that AI hiring algorithms are designed to be unbiased, fair, and effective tools for organisations, debunking the notion of careless manipulation for personal gain.

The view that AI is inherently biased also fails to acknowledge the ethical benefits of implementing AI in the hiring process. AI can accomplish this through various means during the entire recruitment process, starting with the screening stage. By programming software to evaluate candidates based on their skills and experience, a uniform set of metrics is applied to assess every application, as the software will deliver the same exact experience to all candidates, unlike a person, who may assess candidates with intentional (e.g.preferring candidates who went to the same school) or unintentional bias (e.g being tired at the end of the day and paying less attention to a candidate’s answers). This approach fosters an equitable environment where all submissions are evaluated on an equal footing.

For instance, recent research showed that game-based assessments are effective recruitment solutions for businesses looking to improve their accessibility and inclusion to neurodivergent candidates. Game-based assessments offer a stimulating way for neurodivergent applicants to showcase their skills, which can be completed in an environment of their choosing. Research has shown that traditional methods of hiring, like face-to-face interviews, are particularly difficult for neurodivergent candidates and act as a barrier to employment.

Businesses must acknowledge that each candidate has a unique set of strengths and capabilities to offer an organisation. Incorporating diverse assessment options allows each candidate to have a fair chance to showcase their strengths to the hiring manager.

Misconception 2: AI is not a long-term solution for hiring

Often people view AI as a tool to catalyse a business function, essentially accelerating the task at hand. While AI certainly advances productivity by automating business processes and tasks, viewing AI as a shortcut overlooks its endless possibilities for the long run. AI in recruitment significantly improves the time to hire, as well as playing a crucial analytical role as a long-term hiring solution. By examining a company’s data on both successful and unsuccessful applicants, AI can identify patterns and trends in hiring decisions. This enables businesses to recognise potential biases in their decision-making processes and take steps to address gaps in diversity and inclusion. In doing so, AI empowers organisations to proactively improve their hiring practices and foster a more inclusive workforce.

Beyond the hiring stage, AI serves as a valuable tool for employee retention. Businesses can input data regarding an employee’s career growth and progress prior to appraisals and promotion opportunities, allowing for a fair assessment and review of existing employees. By having accurate data that objectively evaluates employee performance, a level playing field is created for promotion opportunities. This not only enhances employee satisfaction but also contributes to improved retention rates within the organisation.

Misconception 3: AI is not effective for decision-making

Another misunderstanding about AI in hiring surrounds the debate about its effectiveness in making better hiring decisions in comparison to a human. Often, this debate is reduced to a simple pros and cons list, weighing the merits of humans versus AI. However, this oversimplification fails to consider the subjective nature of “better hiring decisions” and overlooks a crucial understanding of how AI operates and what it aims to accomplish. AI should be seen as a supportive tool that promotes equity, diversity, and inclusion.

AI offers a much-needed objective measurement in recruitment to search impartially for skills and attributes required for a job role throughout the recruitment process. This distinct capability is difficult for humans to attain as unconscious biases tend to influence their decision-making. Hiring managers often develop preferences and prejudices that are beyond their control, making it nearly impossible for them to achieve genuine impartiality throughout the recruitment process. However, human decision-making can assess candidates on a nuanced level that AI cannot, allowing for interactions in the hiring process such as small talk in the interview and candidate-employer connection that reveals more about a candidate.

On the flip side, it is these interactions where preference in hiring decisions can creep in, opening up the possibility of unconscious bias. Whether that be unknowingly displaying favouritism or discrimination towards candidates due to similarities in background, experiences, or personal characteristics. Here too we have fallen into the trap of outlining the pros and cons of AI versus humans in decision-making. In conclusion, use both. Take the merits of human intuition and connection, and add the objectivity of AI for a blended approach.

In other words, implement AI into hiring to enhance human decision-making, rather than replacing it completely. While AI proves effective in assessing workplace competencies like communication, teamwork, and problem-solving, human involvement remains crucial. Hiring managers need to ensure that candidates align with the organisational culture, and this necessitates human intervention. This approach enables every business to recruit the most exceptional talent that aligns with their organisational requirements and promotes the overall well-being of their workforce.

__

Colin Willis, IO Psychology Program Manager at HireVue.