ChatGPT and human resources: What HR leaders should know

4 HR tasks that ChatGPT can carry out

Some HR tasks are more of a natural fit for ChatGPT than others. Here are some of the areas worth exploring.

Dan Kaplan,
senior client partner
in the CHRO practice
at Korn Ferry

1. Job interview improvements

Helping compose job descriptions is one area in which ChatGPT can potentially save recruiters time.

The tool can also compose interview questions as well as share optimum responses with recruiters, Kaplan said.

Kaplan sees this capability as an improvement over traditional interview tools.

"One of the weaknesses that has always existed with interview guides is [they] tell people what questions to ask [but they] don't tell them what answers they're looking for," he said. "ChatGPT can actually tell you, here are examples of what the right answer will look like."

2. HR content creation

ChatGPT can help HR staff compose text such as an email to send out to employees with company news.

The tool can also add industry-specific context to company communications, said Ben Eubanks, chief research officer at Lighthouse Research & Advisory, a marketing research firm based in Huntsville, Ala.

Ben Eubanks,
chief research officer at
Lighthouse Research & Advisory

For example, HR staff working for an automotive company could ask ChatGPT to add automotive examples when composing an email to staff about benefits changes.

"It [could] say, 'You know how this part on a car works. That's how your deductible works for your insurance,'" Eubanks said. "It will make it [easier] for that audience to understand, and they [won't] need to have all of the nerdy HR knowledge of how all the benefits work."

3. 360 feedback facilitation

ChatGPT can provide a starting point for the creation of 360 feedback.

For example, a manager can prompt the tool to write a 300-word 360 feedback report on a specific individual, then the manager can fine-tune the report, said Bret Greenstein, partner and generative AI leader at PwC.

Greenstein believes this can save managers time and in some cases might even help motivate the manager to create the report.

"It encourages [the sharing of] feedback [when ChatGPT] makes people faster at doing those things," he said. "When it's tedious, you tend not to want to do it because it takes a giant chunk of time."

3. Individualized career pathing

One common HR challenge is the creation of personalized career paths that direct employees toward relevant skills development and training opportunities.

ChatGPT can potentially help with this, said Anthony Abbatiello, partner and workforce transformation leader at PwC. For example, an employee could share their current skills along with their future career goals, such as moving from an associate-level position into a leadership role. The generative AI tool would then carry out the following tasks:

Anthony Abbatiello,
workforce transformation
leader at PwC

• List the skills required for the leadership position.
• Identify the individual's potential skills and knowledge gaps.
• List available learning opportunities that could help fill those gaps.

"The customer-facing functions are really the ones that are generating the best value today in GenAI [if] you look at the employee as the customer of the organization," Abbatiello said.

4 potential ChatGPT HR concerns

As HR leaders potentially apply ChatGPT to some HR tasks, they should be aware of some possible drawbacks as well. Here's more about the possible problems.

1. Lack of sourcing

ChatGPT might use incorrect sources and generate incorrect conclusions, called hallucinations. This, in turn, is likely to negatively affect employees who are using the technology.

HR leaders must be vigilant and ensure that they verify the information they receive from AI tools, Eubanks said.

"[One issue] we could run into is if we're making decisions based on what we see without validating [information]," Eubanks said.

2. Issues with publicly available generative AI

Bret Greenstein,
generative AI leader
at PwC

Generative AI tools that are meant specifically for a certain organization will likely be more helpful for enterprises than publicly available generative AI tools. An enterprise version of ChatGPT debuted in 2023.

Publicly available generative AI might produce vague or inaccurate results, Greenstein said. In contrast, generative AI that sits inside a company's firewall and draws from that organization's enterprise-specific data will be more effective for workplace applications.

"You get an answer that's specific to your business," Greenstein said.

3. Potential lack of employee data literacy

Employees must still possess a certain level of data literacy to correctly interpret data from ChatGPT.

HR leaders must cultivate data literacy throughout their organization and encourage other leaders to do the same, Kaplan said.

"[You need to make] sure that you are doing data science training across your company and building a data-literate employee base," Kaplan said.

4. Lackluster user experience

Generative AI must also provide employees with a positive user experience, or employees will likely not want to use it. Delays in responses or other issues could hinder adoption.

Companies must avoid what Kaplan calls the "credit card company experience," in which callers grow increasingly frustrated when the automated system at the other end doesn't understand their repeated voice prompts. He believes users expect chatbots to quickly provide answers.

"If it's something that causes people to start screaming at their computer because the bot's too slow, you're probably not ready to use that process," Kaplan said.

Carolyn Heinze is a Paris-based freelance writer. She covers several technology and business areas, including HR software and sustainability.