Generative AI burst onto the scene in November 2022 with the potential to automate and improve many aspects of the hiring process, including interviewing. Meanwhile, there’s a rapidly growing movement in business and government to assess skills and eliminate arbitrary degree requirements when making hiring decisions. As a result, for HR and hiring, it appears we are approaching one of those moments in time when powerful trends converge to transform the industry, dividing it into before and after.
But we’re not there yet. HR must do a great deal of work to enable skills-based hiring, and while AI seems as if it could help usher it in, there’s much we don’t know about how HR can safely and effectively use AI. The possibilities, though, are tantalizing.
Skills-Based Hiring
To understand how AI could assist with hiring, we need to understand how the movement to tear the paper ceiling is changing the way in which organizations choose the candidates they want to bring on board. Across the public and the private sectors, leaders are coming to realize that arbitrary degree requirements unfairly and unnecessarily block some candidates who possess the right skills from securing jobs in which they would excel. There’s plenty of evidence that degrees are a poor predictor of performance, and, especially in the technology sector, there are many examples of visionaries who have succeeded without completing a college degree. Mark Zuckerberg, Michael Dell, Steve Jobs and Bill Gates all dropped out of college to pursue their entrepreneurial dreams.
Additionally, given how quickly the world is changing, some skills that were absolutely necessary only a few years ago are already irrelevant. That’s what Gartner discovered when the analyst firm examined job postings from 2017 to 2021. They found that not only has the average number of skills required for a job grown 10% every year, but also one-third of the skills required in 2017 sales, finance and IT postings were no longer relevant four years later. In this environment, college degree programs may struggle to keep up with all of the skills graduates need when these are changing so quickly.
Fortunately, when it comes to hard skills — things like programming languages, new knowledge, regulations, compliance procedures, and the like — self-starters can find resources to fill gaps in their skill sets. Hard skills, however, are not always – or even often – the most important for long-term job success. Organizations should place a high priority on identifying candidates with so-called “soft” or more “foundational” skills that have been shown to be important predictors of job performance – things like critical thinking, conscientiousness, communication, agility and resilience, creativity and teamwork. These are more enduring characteristics that are harder to train than hard skills, but they enable people to learn new technical skills, solve problems, and successfully negotiate complex and constantly changing work environments.
As talent acquisition functions and hiring managers move away from arbitrary degree requirements to skills-based hiring, hiring managers will need to directly assess a variety of critical skills and qualifications that are needed for job success. Therein lies the challenge: ensuring the right assessment tools are in place to evaluate these different skills accurately and fairly, and then scaling these tools to meet the organization’s hiring needs.
Certainly, the use of traditional standardized assessments is one option for doing this, but these aren’t the only – or always the most effective and practical – option. Interviews provide a flexible assessment method that can be used to efficiently assess a wide variety of skills that are needed on the job. but to effectively predict job success, these interviews need to be carefully developed and executed.
Structured Interviews
Interviews can accurately identify qualified job candidates if they are conducted in a rigorous, structured way. Freeform interviews aren’t just non-predictive — they can also be harmful, both to the candidate and the hiring organization. Research has shown, however, that structured interviews are the strongest hiring practice when it comes to predictive power.
Best practice requires that structured interviews have the following:
- Interviewers who are trained on how to properly conduct a structured interview
- Standard, open-ended questions that are the same for all candidates
- Questions that ask about how they applied their skills to past experiences and the resulting outcomes
- Independent ratings of candidates’ responses by a panel, followed by a consensus process to arrive at a final interview score for each
The more organizations adhere to these best practices, the stronger the interviews’ predictive power will be – but implementation requires more time and effort than the typical, looser interview format.
How Can AI Help?
As organizations eliminate degree requirements to focus on skills, keeping up with the increasing demand for accurate and scalable skills assessment will become a significant challenge. AI has the potential to help. For instance, well-trained generative AI could assist with creating and evaluating assessments, and for interviews, it could create well-formed questions, suggest follow-up questions in real time and assist with interview evaluation and scoring.
However, generative AI is a new and disruptive technology, with very little research on how it can be safely and effectively used in hiring, and there are significant risks in using the technology as it currently stands. Prior to wholesale adoption, we need to ensure that there are appropriate guardrails to prevent rouge applications that yield poor outcomes. We need to ensure the AI tools are devoid of biases and that they are in fact accurately identifying the best candidates for a job in fairness to candidates and the organizations using these tools.
Generative AI needs to be trained for the specific HR task at hand. Assessment questions, just like interview questions, need to be properly structured to be effective, and without that specialized training, AI is likely to produce poor assessments and structured interviews that aren’t predictive. Additionally, generative AI must be closely monitored because it can pick up all-too human biases against different demographic groups and people with disabilities, or even bizarre arbitrary preferences such as favoring candidates with specific names.
Then there is the issue of how job applicants will react to the use of AI tools in hiring: Will this turn off the very people the organization is trying to attract? According to a 2023 Pew Research Center study, two-thirds of adults would not want to apply for a job where AI is used to make hiring decisions. Their top worry was that AI would miss the “human” factor, followed by concerns that AI makes mistakes and has design flaws. Companies that use AI risk driving qualified candidates away.
To use AI in hiring, the industry needs rigorous research on the following:
- How to best train AI to eliminate bias and provide well-formed questions for assessments and interviews: Generative AI that is improperly trained on biased data and bad assessments will not produce content that’s useful for HR.
- Understanding what people will tolerate: Many people would never take an interview that they knew was conducted by an AI, for example, though they might take an assessment created by one.
- How to set appropriate guardrails so AI follows professional, ethical and legal guidelines: No matter how efficient AI may be, it won’t be worth it if legal trouble results.
This research is currently underway, and until we learn more about how to best use generative AI in hiring, the wisest approach right now is to wait and see. AI holds great potential for improving the effectiveness and efficiency of hiring, but research-based best practices need to be in place before HR can safely adopt its use.
