By Dr. Shrikant Londhe, Dept. of Chemistry
History has shown that when new technologies emerge, there is often resistance and skepticism regarding their adoption. The introduction of tractors and other mechanized equipment received apprehension from farmers in India, who were concerned about their jobs being taken away by machines. However, those who embraced the technology eventually reaped the benefits in terms of increased productivity and efficiency. In today’s world, educators across all disciplines are faced with a similar conundrum with the advent of Artificial Intelligence in our classrooms and beyond.
To put it into context, we can look at AI as an extension of human intelligence. Human evolution has been characterized by the relentless pursuit of knowledge and innovation. AI, in many ways, is a natural progression in this journey. It symbolizes our collective effort to extend the boundaries of what we can achieve, leveraging technology to enhance our cognitive reach. Common tasks that AI can execute include speech recognition, problem-solving, learning, and decision-making – all hallmarks of human intelligence. AI has reached a stage where it can learn from data, recognize patterns, and even create new patterns through generative algorithms. This marks a significant step in the ongoing evolution of intelligence, as machines demonstrate the ability to go beyond mere mimicry and contribute creatively to problem-solving. Unsurprisingly, AI has grown exponentially in recent years, becoming an integral part of various industries, such as healthcare, finance, and education.
There are plenty of AI based tools available for use to the public. One such tool called Gamma was the subject of my recent exploration. Gamma took my minimally worded prompt about course learning outcomes and provided me with an artistic and professional deck of slides, filled with a decent amount of usable information and images. The generated outcome was only as good as the generic prompt that I initially provided. At the end of the day, I had a framework to take and edit as I please, adding the relevant information as needed. The next obvious task for me was to see how Gamma performed in creating subject-specific content that I can potentially use to teach a class. Although the slides appeared very professional with eye catching designs, the content left much to be desired. Even after providing specific prompts to the AI assistant, the end product had surface level information with no examples, often vague, and misleading in a few instances. This tells me that my intervention is crucial to producing impactful and accurate content even when using AI to create it.
Earlier this year, a user on Reddit posted their experience with generative AI using Bing.1 The user provided a simple prompt to which an incorrect answer was provided by the AI assistant. Upon repeated probing into the issue by giving specific prompts, the AI assistant threw its hands up and excused itself out of the conversation. This example goes to show how AI is very much like us – sometimes it does things incorrectly, gives up in an escalating conflict, and accepts its flaws with grace. It feels weird to personify a digital program in this way, but this is the world that we are stepping into.
A recent Nature article2 discusses how an AI-based program was able to study the inputs given to it, use pattern recognition, and eventually help a human user to decide whether a given write up was generated by a human being or by ChatGPT. A chemistry professor at the University of Kansas co-developed a specialized detector that can catch ChatGPT-generated content in research articles within a specific area of study with near perfect accuracy. Thus, it appears that AI can potentially be employed to tackle problems that we anticipate facing because of the widespread usage of AI.
At the end of the day, humans are inherently social – a trait ingrained in us over millions of years of evolution. We have a deep-seated need for human interaction. Even as machines grow increasingly adept at performing human tasks, perhaps even more efficiently than humans in some cases, the fundamental need for real human connection still persists. Although we can’t predict what the far future holds, we can rest assured that our place as educators in the classroom is not being taken by an AI assistant, at least any time soon.
References
- https://www.reddit.com/r/ChatGPT/comments/14gnv5b/bing_chatgpt_too_proud_to_admit_mistake_doubles/
- https://doi.org/10.1038/d41586-023-03479-4
Dr. Shrikant Londhe, an organic chemistry specialist and chemistry instructor at Oregon State University, is dedicated to fostering an enjoyable learning experience for his students. His objective is to instill an appreciation for chemistry in their daily lives.
Editor’s Note: This is the third in a series of guest posts about teaching strategies by members of the Fall ’23 Resilient Teaching Faculty Learning Community facilitated by CTL and UIT-Academic Technologies. The opinions expressed in guest posts are solely those of the authors.
To join the Winter ’24 Resilient Teaching Faculty Learning Community see the call for participation. Applications are due Dec. 11.
Image by Bing Image Creator
Leave a Reply