I-O, I-O, it’s off to work we go: How I-O psychology is driving AI success
The benefits of adopting artificial intelligence (AI) in HR are significant, but there have also been some scary stories. Setting scary stories aside, it’s only natural to be hesitant about bringing in new technologies. However, there is one critical action you can take to support your success. Hint: It’s not about computing power, it’s about people power. Specifically, the power of industrial-organizational (I-O) psychologists.
I’ve worked with I-O psychologists for more than 15 years, and I will admit that it wasn’t a profession I knew a lot about in those early days. Given that I may not be alone, let’s start with a definition: I-O psychologists observe, analyze and interpret human behavior in a professional environment (SIOP). Mostly, they’re all about making work better for workers and organizations.
By combining the human efforts of I-Os with AI capabilities, organizations are able to make higher-quality decisions through data. I-O-provided domain expertise is also beneficial to AI development and deployment, as their knowledge may be as critical to your success in AI for HR.
I spoke with a number of experienced and qualified I-O psychologists to find out why their expertise is beneficial to AI development. Here’s what they said:
Spot and avoid bias
As experts in human behavior at work, I-Os are well placed to look at the input data for your AI application and ensure that it does not create or enhance any bias. Specifically, they know how to spot inappropriate predictors in a selection context. An example of an inappropriate predictor could be the use of a person’s post/zip code as a predictor of absenteeism. It may seem acceptable, but diverse individuals and those in protected groups may be concentrated in specific geographic areas and using location as a predictor could introduce adverse impact.
Stop you from reinventing the wheel
Human behavior at work has been a topic of much research for decades. Therefore, I-O professionals know a lot about what makes people tick, who will make a good hire, and how to get the most from people once they’re employed by your organization. In their training, I-O psychologists learn about this academic research and can draw on these insights to ensure an AI solution is not wasting time investigating a relationship that has already been proved or disproved. For example, you may think that the number of years of education would be a good indicator of future job performance, but research has shown that general mental ability (GMA) or job simulations are much better predictors.
By practicing scientific research methods, I-Os possess a laser-like focus on measurement accuracy. They understand how to interpret people data and can help you know what exactly you’re measuring, explain the relationships in your people data, and provide information on why your AI algorithms are behaving in a particular way. For instance, an AI solution that offers a measure of a job candidate’s cognitive ability must work effectively and I-Os know that developing an effective intelligence test is about two things:
- Reliability – does it give one person the same score today and tomorrow?
- Validity – does it measure what you want it to measure (in this case, intelligence)?
I-Os identify what’s a real difference and what could be measurement bias. That’s not something machine learning experts tend to address.
Support change management
Implementing AI solutions brings a need for a certain amount of change management. Having a deep understanding of people behavior gives I-O psychologists a valuable perspective on how the deployment is likely to play out in your organization. It may not be a perspective that’s exclusive to I-Os, but they’ve had the formal training in people behavior that can help you empower your workforce so that you get the most out of your new AI technology.
Bringing I-Os to work together with computer scientists is one of the best ways to support your AI in HR success. The skills I-Os have to offer are essential to optimal AI development and deployment. Moreover, given the real interest in AI solutions for HR, it’s time to get busy. I-O, I-O, it’s off to work we go!