There is a lot of buzz regarding uses of Artificial Intelligence today and a growing chorus of voices both advocating and opposing using AI in Education. I happen to be one of those actively encouraging adoption with the right guard rails in place. In this post I’ll share some of my own history with AI/ML technologies, some emerging use cases, and cautions to keep in mind as we move forward with this promising technology.
My first experience with Machine Learning (ML) was working with one of my customers, an Aluminum manufacturer, in the early 1990’s. Surprisingly, Aluminum smelting pots are highly sensitive to small variations in inputs. We deployed an early software program that relied on expert decisions from one of their most skilled employees and used an ML algorithm to adapt these inputs to other smelting pots in the plant. The result was an immediate improvement of over 5% in yield with no other changes. While 5% may not sound like a lot, with the capital and resource intensive nature of aluminum smelting, it was a big deal at the time. This showed me the power of computers to augment and build upon human expertise.
The next leap forward for me was in the early 2010’s. I had the opportunity to beta test and participate in training one version of IBM’s Watson deemed Chef Watson! I love to cook, and had a great time using the early version to generate new recipes and let Chef Watson know that “No, you don’t eat whole bay leaves, you need to take them out of the soup before you serve it!”. This experience clearly showed that AI was best applied to domain specific tasks. It relies on enormous amounts of data on a specific subject and requires careful, consistent feedback to become proficient.
Fast forward to today. Generative AI and especially Large Language Models (LLMs) have leaped to center stage. Each successive advance from neural networks, to natural language processing to generative AI has brought us one step closer to AGI, Artificial General Intelligence. But as Edlyn Levine, Research Associate at Harvard and co-founder of America’s Frontier Fund, states in her opinion piece in Communications of the ACM, 09/23 publication, “LLMs exhibit many behaviors and precepts indicative of intelligence, but are missing something essential: the stuffy rigor of scientific inquiry. AI models are missing the ability to reason abstractly…”
So why is there so much buzz about the advances from LLMs today? In short, ChatGPT ignited a conversation with the masses because it was made available and easy to use for all. I believe this awareness and discussion will help us imagine ways to use it in our daily lives.
In my role as Oracle’s Industry Executive Director for Education I get to meet with a broad spectrum of people doing just that. Alex Feltus is a professor of bioinformatics at Clemson and the founder of an AI based company. He is pushing boundaries by using an AI based teaching assistant in his classes. In a conversation recently, he relayed a story of one of his students and how she gained a much deeper understanding of a python program by running it thru the AI. The tutor was able to explain her own code in ways that made it much easier to comprehend. The future of classroom delivery can be greatly enhanced thru great professors like Alex and the TA he has created.
Another practical example is the Virginia Community College System’s implementation of Oracle Digital Assistant. They are using the AI technology to provide a federated platform across the system that is personalized for each institution. Nicole Stewart, Assistant Vice Chancellor of Information Technology will be sharing details at an upcoming Oracle CloudWorld session and at Educause this fall.
Like any new technology, we need to ensure that our skills keep up. That is where Nancy Ruzycky, a professor at University of Florida’s Herbert Wertheim College of Engineering, comes in. She has been teaming with fellow professor Christina Gardner-McCune to develop Florida’s statewide curriculum called AI Foundations. In addition, she developed the Data Science Frameworks for the state. Both programs benefit from the free Oracle Academy Resources for students and teachers.
New use cases emerge each day and with them we can see exciting possibilities. So now is the time for some cautious advice. We should neither be afraid nor enamored with the technology. Let’s acknowledge that careful training, vigilance against systemic bias, and a trusted enterprise class platform will go a long way to healthy adoption of Artificial Intelligence.
Participate in the dialogue. Share some of your thoughts and questions in the comment section. Diversity of viewpoints will benefit us all.