AI is an exciting, evolving frontier in technology that most people are sick and tired of hearing about. It can feel like AI is the catch-all solution for all of life’s problems, including your professional problems. There are myriad claims about what AI can do for HR management and talent management professionals, including how it can correct bias and discrimination in the workplace.
As in most cases, there is some distance between the promises being made about AI and what it can, or should, deliver to businesses. To shed some light on the upper (and lower) limits of what to expect from AI, we’ve brought together two perspectives on the topic. On one hand, we have an expert on diversity and inclusion (D&I) in the workplace, Dr. Alicia Ingersoll. She has experience building Black Diamond’s Corporate Social Responsibility program from the ground up, as well as conducting research on promoting diversity within businesses’ leadership as a professor at Utah State University.
We also surveyed the TrustRadius community of talent management and HR users, and asked them about the trends in their fields they like and hated the most. Unsurprisingly, AI was among the most commonly mentioned trend, but there was notable disagreement about how reviewers felt about it. Both Dr. Ingersoll and the TrustRadius concurred that AI and its capabilities are ultimately a tool, subject to the knowledge, skills, and biases of the professionals tasked with using it.
While AI might be exaggerated, there’s a good reason it’s been a tech trend with staying power. Dr. Ingersoll argues that AI can play a powerful role in identifying D&I issues within an organization, as well as mitigating bias in existing processes. For instance, various algorithms (the juice behind much of the AI craze) can help mitigate bias in assessments like cognitive testing in the recruitment process. It can also help identify trends of bias internally, such as patterns in performance review data that can be red flags for biased reviewing and promoting.
Current HR professionals we surveyed primarily pointed to AI’s impact on HR reporting as its greatest benefit. Due to the centralization and consolidation of HR processes and dataflows onto unified platforms, HR departments have better access to more data than ever before. Algorithm-powered reporting tools provide the additional horsepower necessary to process and glean actionable insights from the mountain of data. The operationalization of AI in reporting has fueled much of the rise of workforce analytics, or “people analytics,” ranging from “predicting an employee’s tendency to leave their job [to] an application that rates candidates based on their experience.”
In other words, AI can enable HR departments to use data more “intelligently”, and with more awareness of how bias could enter decision-making. However, there are limits to how intelligently algorithms can operate. If not deployed correctly, AI can be underused at best, and perpetuate bias in processes at worst.
The main flaw of AI in HR management is that it is subject to the people who create the algorithm, or who populate it with data. Algorithms get “smart” by learning from preexisting data, and determining trends, patterns, and behavior from that data. According to Dr. Ingersoll, if algorithms learn from biased data, such as certain demographics not getting promoted or hired, or receiving poor performance reviews, an algorithm can easy pick up and “learn” that same bias, since algorithms can’t necessary differentiate between correlation and causation. AI can even pick up bias from the developers coding the original program.
HR professionals we surveyed were also cautious about how much of a blank check should be written for AI capabilities, albeit with a closer focus on how it impacts their day-to-day lives. For some users, AI as a concept has become so vague and bloated that it feels more like a buzzword than a technological concept that’s “better defined and controlled.”
For other users, the difficulty of AI is not its capabilities, but rather the abilities of the people tasked with using AI-powered software. One user cautions that deploying AI must be accompanied by businesses “providing workers the skills to adapt to the changes” that come with the deployment. Otherwise, unprepared HR pros could end up spending more time grappling with a tool too overpowered for their skillset.
Regardless of how you may feel about AI as a trend, it will become an inevitable part of HR professionals’ lives. AI is quickly becoming a core part of HR analytics, and as Dr. Ingersoll attests, if businesses aren’t using HR analytics platforms, “you’re behind the curve.” HR isn’t solely a soft-skills profession, but has adopted the analytical approach that is sweeping most professions.
The ability for AI to be more of a benefit than a hindrance comes down to you. If, and when, you use algorithm-based programs, you have to be aware of how, and why, decisions are being made, and come up with objective criteria to drive decisions– Dr. Ingersoll points to this as a key aspect of addressing bias in decision-making.
Ultimately, AI can’t fix bias, or make better business decisions on its on. “People are complex,” as Dr. Ingersoll points out. “You need other people to address and correct the problem.”