Are You Already Behind If You’re Not Using AI?

“If you spend decades of your career building expertise, it’s scary that a robot can completely replace you and what you do today.” 

That’s a sentiment you’ve probably heard about AI before in almost any industry about any department or role. This particular quote, however, came from this week’s expert guest on America Back to Work, Lydia Wu, about the influx of AI into HR in the past year. 

And, it is scary—especially since human capital management is a path that requires years of experience-building, coordinated professional steps, specialization training, and certifications to build a solid career in HR.

Just take it from Wu, who is in the midst of building that career. She has more than ten years of experience in human capital management and is currently working as a senior director of people strategy and operations at Panasonic Energy of North America, where she is a trusted partner to the C-suite. Wu is skilled at business modeling and case development, change management, talent and people analytics, and HR technology vendor assessment, selection, and management. She holds the SHRM-SCP and Prosci certifications and speaks three languages.

But, in reality, Wu does not believe that AI can replace all of the above—most notably the intangible, human element of a human-centric profession. She does, however, think the new technology will be incredibly disruptive—with the “disruptive power that the internet had when it first came to market”—and HR professionals need to get ahead of it before it reaches mass consumer adoption. 

How AI Is Being Used In HR Now

According to Wu, HR has actually been using AI for the better part of five to six years, largely for two specific use cases. The first is via HR chatbots that help guide candidates through the hiring process or field employee queries about HR-related topics. HR trains or “copilots” those chatbots through AI–through customized algorithms. 

The second major use case for AI in HR today is compensation and total rewards planning, which involves a ton of complicated calculations, mathematics, and human perception. AI funnels all of those inputs into the best possible outcome to help HR pros get to a data-driven answer more efficiently and effectively. 

“Just because we haven’t been made aware of it, doesn’t mean that companies aren’t using it,” shared Wu on the episode. Now, however, with ChatGPT bringing AI into the public eye, public scrutiny of the technology is increasing, and, accordingly, laws and regulations are expected to tighten, too. 

HR Waiting on the Law to Catch Up

Until then, HR is still very much in the early-adopter, pilot stages of AI. Employers are largely in a holding pattern when it comes to compliance and governance since U.S. legislation around AI in the workplace (and beyond) remains nascent—especially compared to the EU with its trailblazing General Data Protection Regulation (GDPR) for mass data privacy and the recently-passed EU AI act. 

“I would argue that there’s a lot of us in the industry who are playing around with A.I. on a smaller scale from a pilot or beta testing perspective,” said Wu. “We’re still waiting for more regulatory signals to help us draw the sandbox in terms of how far is too far and what is safe.” 

Despite the risk of moving too quickly, there’s also a major risk in waiting.  

A Society for Human Research Management survey last year suggested that 79 percent of AI’s use in the workplace is already focused on hiring and recruitment. Another recent study found that the majority of the 250 HR leaders surveyed said they are already using AI across HR functions like employee records management (78%), payroll processing and benefits administration (77%), performance management (72%), and onboarding new employees (69%). 

In order to keep the company competitive, HR professionals need to modernize and prepare the organization for a future with (compliant) AI. Thankfully, according to Wu, there are smart ways to go about doing that. 

Build a Solid Foundation: Clean Up Your Data 

“Start at your foundation. Start at your data technology stack,” urges Wu. 

AI only works when it’s fed with a massive volume of data, training it to come up with its own logic and reaction to certain events. If the data is flawed, too lean, or biased, AI will not “whip up some magic” and get to the right answer. The inputs directly affect the outputs, so companies should start by cleaning up their data and figuring out the right inputs to get to where they want to be. 

Take talent acquisition, for example. If HR is not collecting data at every stage of the hiring funnel to understand what recruiting channels, branding sources, and advertising mechanisms are most effective and efficient for recruiting, AI will not be able to pinpoint where to go looking for the best software engineer or marketing professional—since that data does not exist in the algorithmic environment.  

“The adoption of AI forces us as an industry to look back at the last decade of work that we have done in technology and ask ourselves some really hard questions: did we build our infrastructure right? Do we need to reinforce that infrastructure? Are we collecting the data that we absolutely need for the future?” 

On that note, one of the primary concerns with AI in HR is around bias, ethics, and discrimination. As part of this data cleanup and data infrastructure exercise, HR should scrutinize the data used to train the AI to ensure it’s as fair as possible. 

Choose the Right Tools to Reduce Risk

Some HR teams, particularly bootstrapping teams with fewer resources and younger data programs, are committing to pre-built data models or tools that are built on existing data models specifically for an HR use case to introduce AI into the department. However, using external vendors poses some threats. 

First, employers that use third-party AI vendors can’t be certain that the mass data the model is trained on reflects the company’s internal environment. Second, there are many vendors out there, who claim to have products built on ethical data models, but actually violate privacy standards and fuel discrimination. 

To reduce compliance risk now and in the future, HR departments that choose to go with external, off-the-shelf AI need to vet vendors and their data sets as thoroughly as possible, perhaps teaming up with internal data experts, IT, and legal to get to the best possible decision. 

Organizations with greater resources, on the other hand, are building custom AI models from the ground up. They’re training those models on the hoards of data they’ve collected via their robust tech stacks over the years to avoid the bias and accuracy threats that off-the-shelf vendors pose. Interestingly, however, this method too is falling short at the moment. 

“Even if a company has had a tech system in place collecting data for 20 plus years, sometimes that’s still not enough to train the AI model and get it up to the level of intelligence and maturity needed to make effective decisions.”

While both methods have pros and cons, choosing the right AI tools today is all about aligning company goals with projected compliance responsibilities and present-day company constraints. 

Looking Ahead: Copiloting Is the Name of the Game 

Some predict that in the future, AI may be able to perform sensitive HR tasks like hiring and firing employees or negotiating salary details. In fact, experts predict that by 2024 80% of global 2000 companies will use AI-powered “managers” to hire, fire, and train employees.

But, due to the murky regulatory landscape and the problems with current AI models, that day is a long way off for HR. And, some don’t ever want to see that day come. 

“In its current maturity, I will never argue for a full tech model–where AI is making decisions HR by itself,” said Wu on the show. “The only place I’ve seen successful use cases is where I act as the copilot.” 

According to Wu, AI is here to stay, and it’s expected to shape the future of business, employment, and human resources. That means anyone who hasn’t started testing or tinkering with it in HR is already behind. The best course of action for HR right now is to continue running small beta programs with a strong emphasis on copiloting—where the tool helps drive efficiency on repetitive tasks rather than decision-making. 

Until the regulatory landscape becomes clearer, employers should focus on setting up quality tech infrastructure, and data flows that align with the AI regulations around bias and privacy that are sure to land soon. 

For more expert insights into the current state of HR, AI, and the law, check out this week’s episode of America Back to Work with Lydia Wu. 

LISTEN TO PODCAST