We Asked ChatGPT How It Can Help HR Leaders – Here’s What Happened

HR leaders are busier than ever. Between a competitive job market, layoffs, discussions around remote vs. hybrid vs. return-to-office, recognition & rewards and more, they need all the help they can get. Recently there’s been quite a bit of discussion around AI’s potential in supporting HR professionals, and much of the focus has been on ChatGPT

ChatGPT is the latest iteration of the original generative pretrained transformer (GPT) by OpenAI.  

In simple terms, Forbes describes ChatGPT as a tool that “enables users to ask questions or tell a story, and the bot will respond with relevant, natural-sounding answers and topics. The interface is designed to simulate a human conversation, creating natural engagement with the bot.”

How to use ChatGPT in HR, according to ChatGPT

Is there a better source to determine ChatGPT’s capabilities in HR than ChatGPT?  

If you ask ChatGPT how it can be used in HR, the model generates a response like the following:  
‍

• Recruiting: ChatGPT can be used to create personalized messages to candidates, answer frequently asked questions, and assist in scheduling interviews.
• Employee Onboarding: ChatGPT can be used to provide new employees with information about the company, its culture, and their role, as well as answer any questions they may have.
• Employee Assistance: ChatGPT can be used to provide employees with information about benefits, policies, and procedures, as well as assist with tasks such as booking time off or submitting expenses.
• Performance Reviews: ChatGPT can assist in the creation of performance review templates and tracking employee progress.

• Employee Engagement: ChatGPT can be used to create personalized messages to employees, send out surveys, and track employee engagement.

Of course, you can ask ChatGPT more creative questions like: what should I, an HR professional, ask you? Here's the response we got:

As an HR professional, you may have several questions related to recruitment, employee relations, compensation and benefits, and more. Some questions you could ask me include:

1. What are the latest trends in HR and recruitment?

2. How can I improve the employee onboarding process?

3. What are some effective employee retention strategies?

4. How do I handle difficult employee situations such as conflict resolution or performance management?

5. What are the best practices for compensating employees fairly and equitably?

6. How can I stay compliant with labor laws and regulations?

7. What are some innovative benefits and perks that companies are offering their employees?

8. How can I measure the success of my HR initiatives?

ChatGPT produces accurate, precise, and well-written feedback for any of your industry questions, yet many industry leaders don’t feel threatened.  

In an article for SHRM, Tim Sackett wrote “here's the differentiation between ChatGPT and HR professionals: While you will no longer be needed as the creator of a work product, HR professionals will still be needed to narrate to ChatGPT what they want to create.”

Talking to employees worried about ChatGPT

A recent research article explains some of the shortcomings of earlier GPT models, some of which may help assuage concerned HR team members. [Note that OpenAI intends to troubleshoot with each new version]:  

The shortcut of ‘self-attention’

All communicative AI is reliant on the ‘human labor of understanding.’ In other words, tools like ChatGPT may create the illusion of agency, but each response is just an extraction of something we, the people, have already said.  

In this way, ChatGPT understands and responds to prompts by pulling from the database–human conversations–it’s made up of.  

This is important for HR professionals to flag as concerns about AI take-over rise. Yes, ChatGPT is a powerful tool we can use to save time and act on strategy, but it will never be able to replace the source of its own “intelligence.” Functions like HR will always be uniquely human at its core.  

“Human sense-making is constantly required to keep up the appearances of autonomous or semi-autonomous machine behaviour” - Artificial intelligence and the affective labour of understanding: The intimate moderation of a language model

Allocational and representational harms

When an automated system gives credit unfairly, for example, this is considered allocational harm. On the other hand, representational harm occurs when systems misrepresent or fail to acknowledge groups. For example, InstructGPT and ChatGPT offer vastly different histories about Christopher Columbus.  

If you ask ChatGPT how it can assist HR leaders, it may respond with an answer like this: ChatGPT can offer advice on diversity, equity, and inclusion initiatives. Relying on a system with entrenched biases to develop DEIB initiatives is counterintuitive enough, these initiatives need to be taken on thoughtfully and with empathy.

Monitoring your DEIB initiatives, on the other hand, shows a dedication to the process.  

Small and large worlds

‘The communicative performance of an AI system in a small world is often misconstrued as indicative of its potential performance in a large one’ - Artificial intelligence and the affective labour of understanding: The intimate moderation of a language model  

According to the article, small worlds represent subsets of large worlds. To make a (hopefully) helpful comparison, if a small world is an eBay print of the Mona Lisa, then a large world is the Mona Lisa itself.  

So, even though ChatGPT may perform incredibly well in isolated decision-making situations, that doesn’t guarantee general success.  

Here at Kudos, we’re keeping an eye on all things ChatGPT and how it can further support HR Leaders in building culture and engagement. Subscribe to our newsletter to get the latest HR news and insight direct to your inbox.  

‍