(Note – Areas of the text in italics were either fully generated by ChatGPT in response to my prompts, or alternatively ChatGPT was the primary source of research for the topic.)
Revolutionary technologies abound our world. Hopes and promises of applications or cures for life altering diseases, autonomous vehicles, clean burning engines and more cover the pages of our periodicals and daily news. Yet it always seems that the promise of the technological science is fleeting and that real world application and scale is a distant reality. This fleetingness is particularly felt by the average Jane/Joe consumer who is generally shielded or unfamiliar with large, early development, commercial applications where so many new technologies take root. So too has gone the technological science of Artificial Intelligence (AI).
The concept of AI has been around for centuries. The modern era of AI can be traced back to the 1950’s. In 1956 a group of computer scientists and researchers organized the Dartmouth Conference, where they coined the term “artificial intelligence” and laid the groundwork for AI as a field of study.
The term AI has been in our vocab for decades, but few people could point to a real-world application or benefit of the technology.
Personally, as a supply chain and technology industry professional and now a University Lecturer, I have carefully followed the technological science of AI and Machine Learning for over 20 years. Primarily my focus and interest centered on the evolution of robotic arms and their application on the shop floor of modern manufacturing and distribution facilities.
In November of 2022, through a news release, I learned of OPENAI’s public introduction of “conversational agent” called ChatGPT. Curious, I found my way to chat.openai.com and signed up for an account. Unsure of exactly what a “conversational agent” might be, I began immediately experimenting with questions and prompts to ChatGPT. Early conversations were around “What is ChatGPT?”, and “What can ChatGPT do for me?” I was immediately impressed by the depth and professionalism of the responses. This led me to expand my conversation to areas of personal and emotional interest (for example, what techniques can I use to score more points on the Addams Family pinball machine?) to conversations about helping friends with depression related issues. These early conversations left me amazed at the depth and breadth of intelligence that was readily available. I was reminded of the 2011 CBS TV series “Person of Interest” where an AI like “Machine” was used to help identify national security threats. I asked ChatGPT, what it thought of that comparison, and it replied …
I can tell you that the “Machine” in the TV series Person of Interest is a fictional AI system that uses advanced algorithms and predictive analytics to identify potential threats to national security. While there may be some similarities between my capabilities as a language model and the Machine’s ability to analyze vast amounts of data and provide insights, there are also significant differences in the way we operate. For example, the Machine is depicted as having access to vast amounts of surveillance data and the ability to control various devices, whereas my capabilities are limited to generating text-based responses to user queries. Additionally, the Machine is depicted as having consciousness and self-awareness, which are not attributes of AI language models like myself.
With my interest now completely peeked, I quickly moved on to a professionals view of the platform. As a University Instructor at TCU Neeley School of Business in Fort Worth, TX, I began to ponder how AI platforms such as ChatGPT might change the landscape of higher education? I asked myself, how might I begin to incorporate this incredible resource into my course curriculum and challenged myself to explore the ChatGPT platform further.
While, initially my queries were based on simple topics as I advanced, I began to explore more complex and open-ended queries, such as, what elements are important to include in the development of a corporate strategy, or asking for strategic advice about the next move I should make in a game of chess. I then queried simple math and then more advanced algebraic questions, and then finally onto to questions that required knowledge not provided in the question, for example, industry estimates of transportation costs per mile.
Here is a synopsis of my exploration and what I learned:
- For the most part ChatGPT derived not only the correct answer, but a very thorough and well written response. However, some answers were not correct, particularly questions that required critical thinking skills or knowledge of recent events. (ChatGPT learning base currently does not extend past calendar year 2021)
- After receiving the initial responses, I was able to ask ChatGPT to repose responses in different formats, for example, “please summarize your response in two sentences” or “can you provide that response in a bullet point format”. In all cases ChatGPT reposed excellent responses. In the case of the chess move request, I asked for a chess board diagram of the suggested move and ChatGPT also successfully produced a chess board diagram and description of the next move.
- I posed several questions to ChatGPT verbatim from one of my course exams. The questions were all multiple choice and of the 10 questions posed, 8 were answered correctly. In addition to the answers, ChatGPT also provided background information about why the platform chose the answer. Examples of the questions included a request to complete a linear regression on a set of data and ChatGPT answered that exactly correct. I also posed a question about how best to evaluate an insourcing vs. outsourcing decision and this was one of the questions it was unable to answer correctly, but its reasoning was quite logical and when I explained the question further, it moved towards the right answer quickly.
- Taking what I learned from posing the questions from my exam, I then returned and gave ChatGPT the entire 34 question exam. ChatGPT gained a passing score on the exam. What was most interesting to me though, was focusing on the missed questions. I learned that many of my questions contained unintended ambiguity and after reading ChatGPT’s reasoning to its wrong answers, in many cases I was able to adjust the question to provide greater clarity and thereafter ChatGPT got the question correct.
- Educators should be aware that all assignments, be it homework, quizzes, exams etc. that use multiple choice, written or short answer responses are likely to be able to be directly posed to ChatGPT and that quality and generally accurate responses will be provided by the platform Educators will need to adapt assignments to ensure that recent knowledge, specific case example reference, or critical thinking is required. Questions short on these characteristics will be subject to an easy use of ChatGPT to gain an often-accurate response. My plan is to now rework all my exams to include these noted characteristics, and to pre-screen them on ChatGPT so that I know that if the question were posed by a student to ChatGPT, the platform won’t be able to answer the question directly. Instead, ChatGPT will provide the student guidance and supporting information that when applied to critical thinking skills, the student will be better prepared to provide the correct answer.
- Like all change in life, you can either stick your head in the ground and ignore it and suffer the consequences or embrace the change and adapt your approach to gain advantage. As students continue to widely discover AI platforms such as ChatGPT, they will use them. AI is not going away. It was an eye opener to me that screamed “time to adapt your approach” and to embrace this new awesome technological advancement. Educators need to set the “guard rails” for students such that they understand that AI platforms are valuable resources, much like encyclopedias in my school days, but not the end all. It will be important that educators define acceptable and not-acceptable usage examples to students and further explain to them why these guardrails exist. AI platforms can be great resources for things like, research, writing critiques, learning languages, and just generally brainstorming topics. On the other hand, students need to understand that plagiarizing AI responses or using the platform to cheat on a quiz or exam is not acceptable behavior and is potentially damaging to the students personal development. The importance of developing critical thinking skills is foundational in student development and while AI platforms can be a useful tool in the development of critical thinking skills, students need to use AI as a resource and not a replacement for developing their own mental capabilities.
About the Author: Inspired by serving others, over a storied career, Mr. Mark C. Layton has become recognized as a highly successful entrepreneur, board director, executive leader, lecturer/instructor, and digital commerce industry pioneer. After retiring from a successful career as a serial entrepreneur and C level executive Mr. Layton remains active as a Instructor/Lecturer of Supply Chain studies, at the TCU Neeley School of Business in Fort Worth, TX. Layton, continues to serve as an independent board director, published author, keynote speaker, and as a mentor to students, young professionals, and budding entrepreneurs. You may contact Mr. Layton at mark.layton@tcu.edu.
Mark C. Layton 