ChatGPT is a popular topic of conversation: what it is, how it works and the ethical considerations around its use. In simplest terms, ChatGPT is an app found at chat.openai.com. You go to the app, which has free and paid versions, ask a question and rapidly receive a humanlike answer collated from massive amounts of information fed into the platform.
Let’s say, for example, you ask the app, “What is ChatGPT?” The answer you get might not be exactly what someone else receives. That’s because the artificial intelligence system considers the context you provided and the user preferences you set.
So, you might wonder, what’s the big deal? The reality is that ChatGPT and other AI applications have pros and cons. In this article, I’ll focus on its use in human resource functions at veterinary practices.
The Benefits of ChatGPT
When used appropriately, ChatGPT can save HR departments time and money. For example, writing job descriptions can be a challenge for anyone. However, ChatGPT can draft a job description, and HR personnel can review and edit it for your practice’s specific needs. Then, you could run it past your attorney.
If new employees often ask specific questions, you could do the same: Use ChatGPT to draft responses. You then edit them and, depending on any legal implications, ask your attorney for a review.
In other words, ChatGPT can be incredibly valuable for preliminary work, but human controls are necessary to realize its actual value.
Academy to Innovate HR, an online education site, shared how ChatGPT can screen job applicants and aggregate information on why employees decided to leave. For starters, AI technology can compare a job description and requirements against applicants’ resumes, reducing the time spent by human screeners and, therefore, saving money.
ChatGPT also can create interview questions, but as with anything generated by an AI system, users should review the output carefully to ensure it meets expectations.
In addition, ChatGPT can help you write offer letters, create the first draft of goals and benchmarks, and so forth. And when it analyzes exit interviews and other departing feedback, it can identify commonalities, which might help reduce turnover.
Beware that as good as AI technologies might be, they can’t always pick up nuances as well as people can.
The Challenges
Despite all its utility, ChatGPT has limitations. For instance, it lacks empathy. No matter how well AI can mimic compassion, it only impersonates the human touch. So, when you want to interact with employees in a supportive way, don’t dash off a ChatGPT-generated message, even if it’s factually accurate. That’s because it might not communicate your feelings optimally.
Additionally, and perhaps most importantly, AI content might be inaccurate. AI “learns” by receiving inputs. It isn’t in the business of fact-checking. You might receive ChatGPT responses that were true two or three years ago but not now. Currently, ChatGPT uses information available up to 2021. If we take labor laws as an example, AI technology might tell you what is true in another state but not yours. Or it might spit out information that was never correct.
Sometimes, the errors are minor. Other times, they come with real consequences, such as when an attorney used ChatGPT for legal research and cited cases that weren’t genuine in the Southern District of New York. The attorney, who said he was new to ChatGPT, was fined $5,000.
ChatGPT doesn’t automatically provide sourcing, making fact-checking more challenging. You can request documentation but don’t take it at face value. Instead, use it to verify information provided by reputable sources.
Ultimately, some initiatives are beyond the scope of today’s AI. For example, you might ask ChatGPT to outline an employee manual or business growth strategy. However, humans want a plan that would work in their world, and they need to carry it out.
Also, remember that nothing in an AI database will replace the care, attention and thoughtful insights that veterinary professionals provide to clients and their pets.
Putting ChatGPT to Use
The more specific you are in a question or prompt, the higher the quality of the response. You can pose follow-up instructions, perhaps by asking for clarification of a statement or a more detailed reply. You also can tell ChatGPT to respond in a specific tone — from formal and professional to casual and conversational.
Once you’ve obtained all the necessary information, ask one or more co-workers to check the output for accuracy, tone and branding. Again, sometimes you might need to run the content past legal counsel.
As an HR professional, you have access to sensitive information. Therefore, be careful not to enter confidential specifics into the ChatGPT system because doing so might violate compliance regulations. Plus, know that because the chatbot “learns” from prompts, what you type could be shared with others.
You’ll also want to check for bias. AI technologies return existing information, so any human biases in the original text can reinforce stereotypes. That result can magnify inequalities when used in HR documentation.
Like any other critical aspect of your business, practice leaders should create policies regarding what is and isn’t permissible when using ChatGPT. Because the technology is new, encourage employees to share their findings. For example, they might talk about ways to leverage ChatGPT and when it doesn’t work well.
ChatGPT is neither good nor bad. It comes with significant benefits and flaws. HR professionals must educate themselves about the potential of AI technology and then decide how to harness it best.
HISTORY LESSON
When asked, “Which company was the first with a human relations department?” ChatGPT responded: “The first company widely recognized as having established a formal Human Relations Department was the National Cash Register (NCR) Corporation, founded by John Henry Patterson in the late 19th century.” A Google search confirmed the answer.