Learning to Learn AI

My ongoing work with various forms of artificial intelligence (AI) has me thinking a lot about the learning process. In my personal experience, I believe sports greatly enhanced my ability and grit to learn, while my time as a student at Mott Community College gave me the confidence and curiosity to keep learning.

I leaned into my capacity to learn while interviewing for my first professional role as a graduate intern at Oakland Community College. When the Director of Institutional Research asked about my qualifications for a research assistant position, I knew I had to sell it. I was a 23-year-old with a Communications degree, and my experience consisted mostly of high school jobs at a golf course and various non-research positions on campus as an undergrad (including the mail room and intramural coordinator). When he asked if I knew spreadsheets, I confidently said, “SURE!” His follow-up question, “What software?” was met with an equally confident “LOTUS 1-2-3” (yes, I’m so old that I was using computers before Microsoft Excel was invented.) While my responses had a confident tone, there was little substance behind them — the only spreadsheet I had ever created was to track my scores as a varsity golfer. He quickly responded, “If you know LOTUS, you’ll know Quattro Pro.”

I spent the next several months poring over the Quattro Pro manual like a page-turning mystery novel to learn what I needed for the job. Three years later, I had completed my master’s degree and landed a job as Director of Institutional Research and Planning at Red Rocks Community College. By that point, spreadsheets had become second nature for me.

I recognize that diving headfirst into the deep end of learning comes naturally to me. It fills my bucket — “Learner” is one of my top five Gallup Strengths. Whether it’s learning a new sport — from baseball to golf, racquetball to pickleball, bowling to cornhole, cross-country and downhill skiing to water skiing — learning stretches my brain to think differently and grow in new ways. These experiences have also helped me confront my inherent paranoia and skepticism with AI and become, as I said in my last post, “productively paranoid.” The more I use AI, the more I find myself using it on my terms.

For example:

• Rather than asking ChatGPT to write an entire speech, welcome remarks, or a PowerPoint presentation — tasks that still feel a bit uncomfortable for me — I ask it to provide five choices to help focus my remarks. I often iterate on one or two of its suggestions and ask for an outline, which I then adapt and refine before writing my own final version.
• I find myself using the ChatGPT app more than Google to search and answer both complex and trivial questions. I sometimes use Perplexity AI to explore more sources, but I can also ask ChatGPT follow-up questions to name its sources when needed.
• I once assigned ChatGPT the role of a Termite and Wood Structure Specialist and asked it to analyze a picture I took of some wood on a fallen tree, and it taught me about powderpost beetles.
• Perhaps my favorite experience thus far has been reproducing special memories. Using the MidJourney AI image generator, I created an image for which no photo exists but is forever etched in my mind. Recently, I delivered a keynote at Mid-Plains Community College in North Platte, Nebraska. I opened my presentation by sharing a story about how, 30 years ago, my wife and I drove past Exit 179 on I-80 during our honeymoon trip to Denver, Colorado, in her 1992 red Saturn coupe. After about 15 to 20 prompts and iterations, it produced the image I envisioned.

While the potential negative impact on humanity — and the sci-fi apocalyptic scenarios coming to life — has secured more real estate in my mind than I’d like, setting those fears aside to create space for learning has allowed me to imagine how to use AI in creative ways that have made me more efficient, effective, and productive. The more I experiment with it, the more useful I find it to be.

I fully acknowledge that my “Learner” strength looks different than it does for others with the same strength, and that there are many combinations and themes within strengths that will manifest differently in each of us. However, given the rapid pace at which AI technology is evolving and the myriad ways it can impact us, I hope that everyone — both professionally and personally — takes the time to reflect on the roles AI is playing in their life and considers how to take agency and ownership of its impact.

Right now, I feel like I’m in my second week on the job, devouring that Quattro Pro manual. My brain is on fire, my heart is racing, and the possibilities and challenges all simultaneously seem endless. I’m running as fast as I can — skinning my knees and picking myself back up again — as I race toward some level of competency with this technology before it outpaces me. If you haven’t already, I encourage you to venture on your own learning journey and develop your AI skills at your own pace. AI is here, and it's moving quickly.

If you have any questions or comments on this post, please contact me directly at presblog@mvcc.edu

More Thoughts on Artificial Intelligence

The pace of AI is accelerating, and it’s a bit exhausting. 

Last spring I wrote a blog post titled, I Am a Prompt Engineer. I was so proud to have completed an 18-module online course to build my prompting skills for Artificial Intelligence (AI). Just three months later, OpenAI launched ChatGPT-4 Omni with conversational prompting. Now I start many of my ChatGPT sessions with, “Please refine the following prompt to maximize your response to…” and my original single-sentence prompt comes back to me as a fully flushed-out three- to four-sentence beauty that is far beyond anything I could have crafted on my own. It writes my prompts for me! Just like that, I’m obsolete and my career as a prompt engineer has ended. Additionally, the conversational power of GPT-4o is pretty incredible. I’ve had multiple 10- to 20-minute conversations with it to gain insight or explore issues and situations in ways that expand my thought processes beyond what I could do if I were to continue talking to myself while thinking things through.

I don’t consider myself “Pro-AI.” Rather, I consider myself “productively paranoid.” AI is happening, and it’s happening so fast that it sometimes makes my head spin. I have fun creating images for presentations, but the pace of development coupled with the scale and variety of uses can be overwhelming. This complexity and pace are why I wanted Todd McLees to join us at the College’s Fall Opening. I think his message — embracing AI to advance human flourishing — is important in these early stages, as it encourages us to create more positive-use cases that will hopefully outnumber the inevitable negative ones. I’ve seen multiple articles about the Manhattan Project costing $23B in current dollars and its creators forming an advisory board to create guardrails minimizing the technology’s potential dangers. In contrast, companies are currently spending $230B (10 times that amount!) on AI, with no advisory board or guardrails whatsoever.

We may not be able to influence what the “Big Tech” companies are doing with AI — they are unleashing it into the wild, and its impact is inevitable, whether we want it or not. If we simply ignore AI or fail to develop good habits with it, we run the risk of succumbing to its downsides, much like how social media’s downsides have negatively influenced society. I suppose that can be an individual choice; however, if we collectively ignore AI, MVCC might struggle to remain relevant. Relevancy is what is driving the paranoia I feel, and I’m trying to use that feeling productively by personally engaging with AI.

Recently, the College Senate and Cabinet collaborated to create and launch an AI task force to develop recommendations for a Board Policy on AI at the request of the MVCC Board of Trustees. We’ll also have a small group participate in a 30-Day Challenge led by Todd McLees to develop useful AI habits through daily engagements over 30 days. These actions are all in an effort to stay connected to the blinding pace of inevitable change that now includes SUNY adding artificial intelligence as a requirement under the information literacy core competency of the general education degree requirements.

I appreciate the innovators and early adopters who are engaging with AI and working to guide and support their peers with what they’re learning. I also value the perspectives of the cautious and concerned among us. I share many of their sentiments and believe it is important to include these voices as we look to find our way through this wave of intense change. However, avoiding AI now would be like avoiding the internet in the 1990s. In my opinion, one of the things that makes MVCC special is the constructive sharing of perspectives on complex and disruptive issues. Rather than avoid the hard stuff, I believe MVCC does an amazing job of engaging in meaningful conversations to understand the forces of change and find ways to adapt and evolve our programs, services, and operations to remain relevant. AI may be our greatest challenge yet, but I’m confident we can figure it out together.

*Note: Not a single word of this post was generated by AI. I claim every word in this post as my own.

If you have any questions or comments on this post, please contact me directly at presblog@mvcc.edu 

Prompt Engineering – It’s a real thing

I’m a Prompt Engineer! Well, that’s an overstatement, but I did successfully complete the course, “Prompt Engineering for ChatGPT.” It served a two-fold purpose for me. First, it provided a valuable opportunity to deepen my knowledge and understanding of artificial intelligence (AI) as I’m curious about how it will continue to disrupt learning and other aspects of our daily lives. Secondly, I found the class on Coursera, the largest Massive Open Online Course (MOOC) provider that offers more than 7,000 courses to more than 100 million students globally (19 million in the U.S. alone) in a variety of languages.

My prompt engineering course was taught by the Computer Science Department Chair at Vanderbilt University, and he was great. I was intrigued by the business model of charging just $49 per month until I completed the course. Initially, I had high hopes of completing the six-module course in one month; however, as is often the case for so many students, life got in the way, and the course took me two months to complete.

Given that most of my formal education occurred prior to the Internet (for real), I never considered myself an “online learner” and stayed away from pursuing any online offerings. However, I’ve been closely observing MOOCs for the past decade, witnessing their remarkable appeal to millions of people around the world. I was pleased with the convenience of blocking out an hour here and there on weekends to watch the video lectures, do the assigned readings, take notes, and complete the assessments.

It was interesting to assign prompt and audience “personas,” such as the role of a travel agent crafting a vacation itinerary, a 7-year-old requesting a bedtime story, a woodpecker learning to survive an Upstate New York winter, or an organization like MVCC reflecting on aspirations and apprehensions regarding its trajectory over the next five years. The practice of submitting prompts and asking ChatGPT to suggest improvements for those prompts also proved to be unexpectedly enlightening. It’s a strategy I hadn’t previously considered, but it sure makes sense and works quite well!

Some assignments were more interesting than others. I enjoyed the Few-Shot prompt pattern where I had to describe the input (such as a student needing to drop and add a class), instruct ChatGPT to “think” (for instance, securing the student identification number), and then direct it to take “action” (such as entering the student ID, securing the class schedule, and executing the drop/add process). With little effort, I essentially surfaced the basic programming behind an AI-enabled scheduling bot. This is why I love the slightly trick question — “What is the most important programming language for AI?” — the answer being “English.” It’s all about phrasing and leveraging the prompts to achieve the desired outcome.

The course went on to teach several different prompt patterns, including Game Play, Template, Recipe, Chain-of-Thought, Ask for Input, and Outline Expansion. My favorite, however, was the “Alternate Approaches” prompt pattern where I provided some context and asked for alternative perspectives or strategies. I thought of an AI career coach and described myself as a young college graduate interested in a 40-year career with distinct goals in a chosen field and asked for two to three job titles in four- to five-year increments throughout the 40 years. ChatGPT came back with four or five questions to clarify my interests, then provided an informative outline of three different career paths complete with relevant job titles to consider at each stage of my fictional career. Fascinating.

I still have a lot to learn about artificial intelligence — we all do. Its impact on our world is already palpable, and I believe it will only continue to accelerate disruption in ways that are hard to comprehend right now. Here’s an interesting opinion piece about AI (it may have the NYTimes firewall) and the importance of human skills in an uncertain future. While some people fear the potential negative possibilities of AI and others are excited about it (including investors in the stock market these days), I’m trying to stay tethered somewhere in the middle to see the pros and cons more clearly. Given all the unknowns with AI, one thing I do know is that I need to take responsibility for my own learning and continue finding ways to “muck around” in the messiness, building my understanding and honing my skills to successfully navigate an AI-enabled future. I hope you do the same.

If you have any questions or comments on this post, please contact me directly at presblog@mvcc.edu