Thinking with Machines (Part 1)

I teach Generative AI to older adults.

In many ways, that simple fact is what led me to the question at the center of this series. As AI tools become increasingly capable and accessible, many older adults are trying to understand what they are, how they work, and whether they are worth using. Some are enthusiastic. Some are skeptical. Most are curious.

I am curious too.

A topic analysis of my blog and Substack writings since 2003 would likely show a long-standing interest in knowledge, learning, and cognition. Those interests have followed me through different stages of life and different careers. Today, I have the luxury of time that I did not have while raising a family and earning a living. I can pursue questions simply because they interest me. I can also spend time with people my age and older who are asking many of the same questions about technology, learning, and the future.

This is not the first time I have encountered a technology surrounded by both excitement and concern.

Earlier in my career, I worked in international development and became deeply involved in discussions about information and communication technologies (ICTs). At the time, many people viewed these technologies as either transformational solutions or overhyped distractions. In 2004, I contributed a chapter to a book that examined ICTs as tools for development. The subtitle of the chapter included a phrase that has stayed with me ever since: between skepticism and optimism.

I find myself returning to that perspective today.

AI offers remarkable possibilities. It is not simply another software application. At the same time, many of the familiar patterns are still present. There is hype. There are exaggerated claims. There are barriers to adoption. There are concerns about unequal access and the possibility of a widening AI divide.

What feels different to me is the nature of the technology itself.

Unlike many previous technologies, AI increasingly participates in activities that we associate with thinking. It can summarize information, generate ideas, answer questions, explain concepts, draft text, and support decision-making. Whether these capabilities ultimately strengthen or weaken human cognition remains an open question.

That question feels particularly important in later life.

Many older adults are already thinking about how to maintain cognitive health, continue learning, and remain independent for as long as possible. Cognitive decline is a concern regardless of whether AI exists. If AI changes the way we think, learn, remember, and make decisions, then understanding those effects becomes especially important for aging populations.

The question is personal as well.

As someone who uses AI regularly, I wonder about its impact on my own cognitive capabilities. As someone who teaches AI to older adults, I feel a responsibility to understand both the opportunities and the risks. How should I think about AI as a learning tool? How should I teach others to use it? Which capabilities should be supported, and which should continue to be actively exercised?

Those questions eventually led me to launch a Learning Sprint focused on cognition.

I have written briefly about Learning Sprints before. A Learning Sprint is a structured and intentional effort to learn about a topic. In practice, it often resembles a rapid prototype of a course, with myself as the intended audience. Most of my learning sprints have a defined timeframe. Four weeks is often enough to reach the level of knowledge saturation I am seeking.

This one turned out to be different.

Cognition had already been an ongoing interest for years. I had accumulated articles, books, research papers, notes, and observations. Rather than beginning with resource collection, the challenge became one of synthesis, analysis, reflection, and writing. The sprint evolved into multiple phases. At this point, it is probably no longer accurate to call it a sprint. It has become more of a slow burn.

The initial question I settled on was this:

How do humans maintain and develop cognitive capability in AI-mediated environments?

That question became the organizing principle for everything that followed.

In the next post, I will explain why much of the literature I encountered seemed preoccupied with what AI might cause us to lose. Terms such as cognitive offloading, cognitive debt, and cognitive decline appeared repeatedly.

Yet the more I read, the more I found myself wondering whether we might also be asking a different question:

How can humans continue to learn, adapt, and remain cognitively capable while thinking alongside increasingly capable machines?