This report aims to give HR and other business leaders a basic, nontechnical foundation in AI and a deeper understanding of what AI is and isn’t. With this knowledge, they can recognize AI hype in the marketplace and realistically assess where AI might make a difference in HR’s capabilities. With a grasp of the fundamentals of AI, HR will become a more valuable partner in thinking through the optimal mix of people and technology to deliver business results.

Executive Summary

Stroll through the exhibition hall at an HR technology conference, and you encounter a vast array of digital solutions, many of which are touted as artificial intelligence (AI). Yet, strictly speaking, few of those products qualify as such, based on our gold-standard definition:

AI is technology that mimics human thinking by making assumptions, learning, reasoning, problem solving, or predicting with a high degree of autonomy.

That’s a definition that AI experts would subscribe to. But in the layperson’s world, the term AI is often applied more loosely to mean the use of computer systems or agents to perform any task that, up until now, humans had to do. The disconnect between the expert’s view and the popular one causes some confusion in the business world. AI is perhaps the hottest topic in business; yet, most of the so-called AI systems and tools that organizations use—including in HR—fall short of the gold standard. Rather than thinking about AI as a binary (i.e., yes/no) concept, therefore, it’s more useful to imagine AI as a range, with assisted intelligence at one end and autonomous intelligence at the other. 

The model below can help HR leaders become better potential customers of AI for HR. Rather than holding out for the gold standard, it presents shades, or gradations, of “AI-ness” that reflect the range of tools in the marketplace. Importantly, the model also captures what’s not AI: automation; that is, the tools and systems that can perform simple, repetitive tasks independently, although they cannot think or learn in the process. Automation is effectively off the AI grid, therefore, but we include it in our model, set apart from AI, because the two are so often conflated.

Insights for What’s Ahead

• While related and often conflated, automation and AI aren’t the same; buyer beware. Automation doesn’t meet our definition of artificial intelligence. It describes a range of tools and systems that can perform simple, repetitive tasks independently but cannot think or learn in the process. Automation is still very useful in enhancing efficiency and productivity of HR processes, but it shouldn’t be confused for AI as it lacks the ability to adapt.
• To mitigate the risks of AI, HR practitioners should look for AI with “explainability,” a basic principle of responsible AI. While AI provides a tantalizing vision of a more productive workforce, it can also bring risks of bias and ethical challenges. HR leaders do not require an advanced understanding of AI to understand what factors are being included in AI-enabled decisions or what data sets the AI is learning from. Being able to “peer into” the black box of AI will go a long way in mitigating the risks of AI.
• HR leaders can adopt AI to significantly enhance the internal customer experience, supporting engagement and retention of key talent. AI can aid HR in personalizing the employee experience, tailoring both programs and outreach based on individual needs and preferences. This personalization can apply across all functions of HR, from triggering automatic nudges to combat implicit bias, to providing networks for new hires, to even suggesting rewards and compensation packages. The potential of AI in HR can be significant.

How HR will work differently in the future with AI

Many HR organizations are making a determined effort to treat employees like customers, enhance employee experience, and customize HR offerings to specific segments and individuals within the workforce. AI can advance those efforts.

In the future, companies may use AI to:

• Assess and rank candidates based on their role fit and readiness to switch jobs. The hiring manager uses this output to make the final selection decision.
• Analyze real-time data from multiple sources and then add or subtract resources from projects, manage deadlines, schedule meetings, take notes, and do basic follow-up.
• Track when new hires have completed required training; point them to coworkers they should connect with, based on their respective skills, experience, and jobs; and recommend other relevant development opportunities.
• Make customized health-benefit recommendations to individual employees, based on their personal profiles, family circumstances, past usage, and other factors.
• Listen in on online sales presentations and provide sellers with real-time, pop-up suggestions to improve their pitches.
• Suggest compensation packages based on performance data, external market trends, and comparable internal data, to ensure managers award equal pay for equal work.
• Nudge managers to avoid referencing stereotypes or using biased language
• in performance reviews.
• Match employees’ skills and experiences to future role openings to widen
• the pool of potentials and increase internal mobility.
• Mine qualitative data, such as online discussions and open-ended polling questions, to track employee sentiment and identify organizational issues that could affect attrition.

The inherent risks of AI for HR

While the potential of AI is enticing, HR leaders must closely consider its inherent risks. These include the “black box” of AI algorithms—the potential bias or error that users cannot see—and the ethical, legal, and fairness concerns that could result. Understanding what AI cannot do is as important as appreciating what it can. Take, for example, its inability to extrapolate from one situation and apply that knowledge to similar, but slightly different, circumstances, as humans can.

That doesn’t mean that HR teams should eschew AI altogether, simply because they don’t understand it or see it as too dangerous. Rather, they ought to proceed with eyes wide open and with appropriate safeguards and controls in place, such as those outlined in the report.