Today’s Tuesday Reading is from Jim Bruce, Senior Fellow and Executive Coach at MOR Associates, and Professor of Electrical Engineering, Emeritus, and CIO, Emeritus, at the Massachusetts Institute of Technology, Cambridge, MA. Jim may be reached at [email protected] or via LinkedIn.
To intelligently use AI, we must first understand how we got here. The name is more recent than the concept. Artificial intelligence is quite old. According to Wikipedia, “The history of artificial intelligence began in antiquity (that is, in BCE time), with myths, stories, and rumors of artificial beings endowed with intelligence of consciousness by master craftsmen. The study of logic and formal reasoning from antiquity to the present led directly to the invention of the programable digital computer in the 1940s, a machine based on abstract mathematical reasoning. This device, and the ideas behind it, inspired scientists to begin discussing the possibility of building an electronic brain.” [1] In a concurrent effort, the Atanasoff-Berry Computer (ABC) was built between 1939 and 1942 at Iowa State College (now University) by John Vincent Atanasoff and Clifford Berry. This machine was designed to solve systems of simultaneous linear equations, which were becoming common in physics. It could handle systems with up to 29 equations. [2]
The “electronic brain” did not arrive until later, after further developments in computer hardware and software and the concepts of artificial intelligence. “The field of AI (artificial intelligence) research was founded at a workshop held on the campus of Dartmouth College in 1956. Attendees at the workshop became the leaders of AI research for decades. Many predicted that machines as intelligent as humans would exist within the next generation. The U.S. government provided millions of dollars with the hope of making this vision come true.” [1]
In the coming years, scientists and their government funders realized that they had grossly underestimated the difficulty of this challenge. As a result, funding decreased, though some research continued, often under the guise of different names. By 2000, “deep learning” was applied to numerous problems in academia and industry. Deep learning is an artificial intelligence approach that teaches computers to process data in a way inspired by the human brain, much like a child learns by example. As a public demonstration of this, in February 2011, IBM’s Watson DeepQA computer defeated the Jeopardy quiz show’s two all-time champions, Brad Rutter and Ken Jennings.
Taking these concepts further, scientists have extended AI to create new content, such as text, images, videos, audio, and music. GenAI (generative artificial intelligence [3]) is a form of artificial intelligence that learns from training data to produce new data as output. I used a GenAI tool, Perplexity, in my preparation for writing this essay.
For example, I asked Perplexy to “give me ten possible topic headings for a newsletter article on AI in leadership.” It provided a list of ten topics from which I selected three and asked that these three be expanded upon, with some detail. Here is a portion of one of the points:
Transforming Leadership Roles: The Impact of AI – Shift from traditional leadership to AI-assisted leadership:
A. AI as a tireless assistant for data analysis and decision-making.
B. Automation of routine tasks, freeing time for more strategic work.
C. Enhanced decision-making capabilities.
D. Challenges in adapting to AI-driven leadership.
E. Emphasis on emotional intelligence and interpersonal skills.
F. Strategic thinking and vision.
G. Ethical responsibility in AI implementation.
H. Personalized leadership development.
Let me pause and ask why we might want to use AI in a college or university. I think that there are several key reasons. [4] First, AI has the potential to improve operational efficiency and effectiveness, improve decision-making, help the organization innovate, and quickly analyze data, thus enabling more informed decision-making. Second, these efficiency improvements will free staff to do more meaningful, creative work. For example, most IT organizations have Help Desks staffed by some number of people. Imagine that the answers to most of the questions they have historically received have been developed and stored in a database so that they can be retrieved when a call with that question is received, answered by AI that analyzes the question and connects the caller to the answer. And the entire transaction could be done either via email or orally on the phone.
Similarly, this same type of help is provided today by staff in most of the administrative offices around the campus. AI could also be used to provide helpful, thorough responses in these situations as well.
AI could also be used to analyze Help Desk data and data from all other calls to an organization, permitting the individuals who would typically answer these questions to do more complex and creative tasks. (Surveys from 2021 have shown that administrative staff spends about 4.5 hours per week on routine, mundane work that could be automated.) Meanwhile, the data could be analyzed for insights and potential recommendations to improve the organization’s services and effectiveness. AI could not only assess quantitative data, such as call volume trends and response times, but also qualitative data, like the sentiment and themes of customer interactions, providing a comprehensive understanding that drives targeted improvements.
Some individuals fear that AI cannot be trusted. This fear often arises from a need for more familiarity with AI models and how they work, whether the rules on which a model is based are complete and unbiased, whether the model has been adequately tested, etc. For example, ChatGPT sometimes has difficulty with complex numerical analysis. This is likely because it is a language model trained on text and not on math. AI can also suffer from hallucinations that create false responses. It is important that all AI-generated content be validated.
Today, a growing number of college and university students are using AI to help them with their schoolwork. They may be using AI to help them brainstorm ideas on a topic of interest. Or they may use an AI tool to outline a topic. Or inappropriately, I think, to write a paper for them. Some universities are reportedly integrating AI with their record-keeping systems for data analysis of various financial and personnel records. Others are beginning to experiment with using AI for personalized learning. Some also express concerns about AI’s reliability and ethical implications in an educational setting, as generative AI systems can lack context and accuracy. [4]
At this point, we need to accept AI as beneficial if used responsibly and learn to use it for positive benefits.
I would love to learn about the use of AI on your respective campuses. So, I invite you to send me a quick note about your campus’ use of AI. If we receive sufficient responses, we’ll share results at some point early in 2025.
If you’re currently unfamiliar with AI, I hope that in the coming weeks, you will take the time and put in the necessary effort to learn more and begin to put it to positive use across the many areas of your life. Have a wonderful, productive week … jim
[And, a hardy thank you to David Bruce, Vice President and CIO at Simmons University, my son, for helping with the example and teaching me how to make AI queries.]
How do you envision AI impacting your work in the next year?
Last week, we asked about your current work goals:
- 45% said they could focus more on the parking lot and less on the summit.
- 28% said they’re good at focusing on the parking lot.
- 27% said they don’t have well-defined goals.
Last week’s metaphor of hiking to the summit vs. making the goal a safe return to the parking lot seemed to resonate, with about half of us identifying the opportunity to focus more on the parking lot and less on the summit. When thinking strategically, it is critical to consider the full picture of our goals and the long-term sustainability of them. And for the roughly 1 in 4 of us without well-defined goals: you can’t make up in tactics that which you lack in strategy. Perhaps the final weeks of 2024 are a good time to lead from where you are and generate greater clarity around goals.
References
- History of Artificial Intelligence, Wikipedia
- The Atanasoff-Berry Computer
- Suriel Arellano, Leading in the Age of AI, 2023.
- A number of generative AI sites are available for public use: ChatGPT (OpenAI), Gemini (Google), Midjourney (for image generation), Copilot (Microsoft), Character.ai, and ElevenLabs (for voice generation).
- Inside HigherEd. Benefits, Challenges, and Simple Use Cases of Artificial Intelligence in Higher Education