ChatGPT, the workplace and HR: Setting the scene

In his 2022 year-end LinkedIn post, my HR colleague and co-author Peter Vermeulen reflected on the trends that currently shape HR and the workplace landscape. He disclosed that he tried out ChatGPT to write that post.

The major trends the chatbot summarized in a few seconds—continued shift towards remote work, focus on diversity, equity, and inclusion, and increased use of technology in HR—and its conclusions—HR will need to be adaptable and proactive in order to meet the evolving needs of their organizations and their employees—both startled and resonated with me.

Just weeks prior to that, I had spent considerable time researching the various HR trend reports to illustrate my own monthly HR column and to prepare for a keynote presentation. My analysis and conclusions were very similar to the chatbot’s summary!

Could the expertise and gut intuition of seasoned HR professionals, honed over a decade of learning and skill building in the trenches, be supplanted so effortlessly by a chatbot?

Rather than following my gut-feel and critical academic inclination to immediately criticize the tool, I decided to spend some time looking into the applications and uses of ChatGPT, and more specifically, its impact on HR and the workplace.

A transformative end to 2022

Since an updated version of the chatbot was released for public use by OpenAI on November 30, 2022, many (HR) influencers and journalists have already shared positive and negative uses and implications of the tool.

My own understanding of the implications of AI on HR has been seeded by two colleagues whose thinking I am not convinced could ever be replaced by technology.

I owe a great deal of gratitude for speeding up my learning curve on AI and HR to Peter through the many conversations we had about HR technology and digitalization of HR services when he was an HR executive at Amazon and we were discussing the content our Be(come) an Awesome Manager book.

The book also includes Peter’s brilliant case study on the use of artificial intelligence in HR entitled, Innovation at Amazon through Culture and HR Technology. My own thinking on the digitalization of HR over the years has also been enhanced by regularly following the HR Tech updates of Josh Bersin.

In terms of the original question: could human judgment and experience be surpassed by AI in the all too human field of Human Resources? At first glance, yes. ChatGPT was obviously much more productive in cycle time to produce the HR trends than my own research efforts (i.e., a few seconds by the AI in contrast to several hours for me).

But are there any other significant differences between the way I researched the HR trends—using traditional academic methodology—and the output of ChatGPT? The chatbot uses a generative AI model, meaning that there are no links to the sources of information used or ways to check the references and whether the information is accurate. I used a semantic human search model (augmented by search engines), retrieved and read the relevant sources, and used my academic experience and human judgment to evaluate their credibility and the information contained.

But let’s not be too arrogant about my cognitive abilities as an academic! I strongly believe that the well-defined corpus of global HR and management knowledge I acquired through years of research, HR experience, and teaching can easily be duplicated, enhanced and even surpassed through AI in the near future —provided there are some checks and balances— and quickly become even more credible in accuracy, trustworthiness, and ease of keeping the state of research in my field updated.

What is ChatGPT?

ChatGPT is a neural network machine learning model—GPT stands for Generative Pre-trained Transformer—using natural language processing (NLP) to assist with language-related tasks such as providing information, answering questions, and generating written and conversational text based on the input of queries by users.

The deep learning-based language model can produce humanlike text using a simple prompt in the form of a question and provides human-like responses in seconds based on the knowledge gained from the data on which it was trained (in this case using the information on the Internet and other data sets).

More specifically, ChatGTP is a text-generation chatbot tool that can generate ideas, draft and troubleshoot text (including code), simplify complicated ideas, summarize content, write essays, and provide answers to a variety of questions designed and automate a series of tasks in a convincingly conversational manner—all requiring minimal training by its users.

While the tool has no inherent intelligence, the AI system is self-learning and can machine learn from its millions of users and from smaller but validated deep databases as it gets bigger.

Used as a starting point, it can generate first drafts of a series of different documents (emails, legal documents, summaries, etc.) which can then be verified and augmented by human touch to produce blended writing.

Since its recent release, the chatbot has generated a great deal of interest because of its wide-ranging applications in the workplace—including HR. A January 2023 survey among 4,500 professionals from various companies conducted by Fishbowl (Glassdoor) indicated that over a quarter of professionals (27%) have already used AI to assist them with work-related tasks.

There are differences in the adoption of AI tools by generation, gender and industry. Gen Z professionals have the highest rate of adoption at work (29%) followed by Gen X (28%), and Millennials (27%.); professional men (30%) used AI significantly more often in the workplace than women (20%); and there are significant differences in usage of AI to assist with work-related tasks by industry with marketing and advertising (37%), technology (35%) and consulting (30%) leading and teaching (19%), accounting (16%), and healthcare (15%) lagging.

Another recent IBM study, indicates that 35% of companies are already actively using artificial intelligence in their day-to-day operations and 42% are still exploring its potential.

Part two of Lisbeth’s essay will focus on use cases and the impact of ChatGPT, and will be published this Friday – in the meantime, you can read more stories about artificial intelligence here.