By now, many have heard of ChatGPT and the efficiency it has provided for many. ChatGPT was created by OpenAI, an AI and research company.  ChatGPT is a chat box which uses natural language processing driven by AI technology in an open environment.  Currently it is free to use, but eventually it will likely be monetized because the actual computing costs are high.  ChatGPT in its current form can provide answers to questions, write emails, compose essays, and write code. 

ChatGPT is not a search engine like Google.  It cannot search the internet.  However, it is a learning algorithm based on training data to generate a response.  Training data, in a simple description, was developed by having human AI trainers conduct conversations in which they played both parts, the user and AI assistants. The response can be wrong. 

Using ChatGPT, students could bypass actual work by having the chatbox write an essay.  It can be done in multiple languages. This is problematic.  Educators are either banning ChatGPT or trying to find ways to identify ChatGPT versus student written essays.  Interestingly enough, Edward Tian, a 22-year-old senior at Princeton University built an app called GPTZero to fight AI plagiarism. It has become so popular that the app has crashed at times. 

How does it know if the text is AI or human generated?  The app uses two approaches:  "perplexity" and "burstiness."  Perplexity measures the complexity of the text.  The more complex and less familiar, the more likely it was written by a human.  Burstiness measures the complexity of the text.  AI tends to be more uniform.  Human will intermix long and short sentences.

So, what positives should HR take away from ChatGPT?  First, it will change training from how to do a job to how to use AI to do a job, and it reduces training time.  "You won't have to train people to do certain things anymore, but rather train them how to use AI tools to do those things," Ricardo Michel Reyes, chief science officer and co-founder of AI company, Erudit says. "For example, before managers would have to train people to make PowerPoints. Now, they won't even need them to know how to use PowerPoint, just understand enough to ask the AI the right questions."

Yet there are some drawbacks.  AI doesn't have free will," Reyes says. "It still needs someone to tell it what the PowerPoint is about, what goes first, and what colors and fonts to use. The problem is that often employers will think that a tool is enough by itself or that they don't need to retrain or retain people. Just as many things can go wrong as things that can be really great. It just depends on the humans and especially the managers — who will need to train people how to use it to be able to profit from it." 

But it does change the possibility of what type of worker to hire.  With AI, the worker may not have to be as highly skilled or knowledgeable.  Therefore, it could open more doors to new workers for jobs previously cut off from them because of credentials. In other words, hiring becomes easier, as long as the training is in place.

On the other hand, ChatGPT is being used to write business emails.  Christian Lanng, CEO of digital supply chain platform Tradeshift, used the platform to write emails and claims no one has noticed the difference. He even had it perform some accounting work.  It could also be used for writing performance reviews, emails, handbooks, and the like.  This situation also leads to a concern that employees might use ChatGPT in connection with work that is confidential and requires data privacy.  Employers are naturally concerned that employees will share proprietary, confidential, and/or trade secret information when having “conversations” with ChatGPT.  HR should stay on top of this technology and use within the organization, and likely with legal counsel, set up parameters when the tool cannot and should not be used.  It’s a new world for HR.

Source: Seyfarth Shaw 2/21/23, CNN 1/24/23, ZDNet 1/23/23, EBN 1/23/23, NPR 1/9/23