Before getting into today’s essay, a brief note about some events I will be doing related to the launch of my new book, Why Health? What We Need to Think About When We Think About Health. On September 24, I will join Jessica Steier, of Unbiased Science, for a Substack live conversation about the book. On October 5, I will join Katelyn Jetelina, of Your Local Epidemiologist, for a conversation in Washington, DC, at Busboys and Poets. On October 22, I will be at The Novel Neighbor in St. Louis, MO, for a reading and Q and A. I look forward to these events, to continuing the conversations that shaped the ideas in Why Health?
Now, on to today’s topic.
This is the second part of a two-part series on AI and the future of knowledge generation. Last week’s essay addressed the intersection of AI and universities, touching on how higher education can survive and thrive in the AI era. Today, I would like to move from what AI means for knowledge generation at the institutional level to what it means for individuals. How should those of us who spend our careers in the idea space approach AI in our own lives and work? What does a healthy engagement with AI look like? What standards and best practices should facilitate this engagement in academia and beyond? These are difficult questions, not to be answered lightly, but rather through a process of conversation, of trial and error, of a collective working out of how to best navigate this technological moment. In the spirit of contributing to this conversation, some thoughts on thinking and writing in the AI era, and what this new technology may mean for those who work in the academic space.
Over the last few years, AI has increasingly dominated the conversation about writing and ideas (as well as conversations about much else). Within academia, AI is the elephant in many rooms, raising questions about what this new technology means for the fundamentals of how we do what we do. Questions like: what does AI mean for the practice of research, peer review, and other core academic tasks? Should AI be used exclusively for technical work, or does it have a role to play in the actual generation of ideas? Perhaps at the more existential level, should we be afraid AI might replace us—replace the individuals and institutions whose job is to think, teach, and engage in the idea space? Is AI a tool, even a partner, to be welcomed, or a rival to be warily regarded?
To think about AI is, in many ways, synonymous with thinking about an advanced technological future, with all this entails. Regardless of the form AI takes, it will likely define the coming era—in ways we can predict and in ways we cannot. I enjoyed recently exploring this a bit by hosting a podcast with Professor Janet Rafner about hybrid human-AI engagement in creative work. It was interesting to look at the many ways human-AI co-creativity can help produce work which is of genuine intellectual and aesthetic value. Such co-creativity has already yielded much and has the potential to be a true driver of human progress in a range of sectors, potential which may fairly be called revolutionary. Yet it is important to remember that revolutions can have messy downsides, destroying as much as they create, upending deeply rooted, necessary things. In the case of human-AI co-creation, this may include the uniqueness of human creativity, its standing as something distinct, something of which nothing outside of our species appears capable.
Given these possibilities, it seems reasonable to approach AI with a measure of curiosity and caution. Curiosity because AI is already doing much to make the world a better place and has the potential to do much more. AI is having a rapidly growing influence in coding, data processing, task automation, medicine, even the arts and humanities. We should want to learn more about all this, to participate in this revolution and to benefit from it. Caution comes when we acknowledge just how fast all this is moving, how unpredictable its outcomes remain, how thoroughly AI has shifted the ground beneath us already, even at this comparatively early stage in its emergence. We are speeding towards a future we do not fully understand, cannot fully envision. This should give us pause, urging us to be reflective on this journey, to take stock of where we are and where we may be going as we acknowledge none of us can predict our destination with complete accuracy.
This suggests a need for balance, for pursuing a thoughtful engagement with AI along with a healthy respect for its disruptive capacities, recognizing that we are in uncharted territory with all the opportunity and potential peril this entails.
How do we strike this balance?
The paradigmatic case that merits conversation here seems to me the use of AI in writing. We might start a consideration of AI’s role in writing by asking: Why do we write at all? We write to understand, to think better. That means that those who are privileged to engage with ideas for a living have a responsibility not to let machines think for us—not to let AI do anything that denies the world the best of what our minds can produce. Writing, after all, is not just the process by which we express our thoughts, it is in large measure the process of thinking itself. We develop ideas by writing, by working through subjects on the page, and we aspire to think better so that, when we express ourselves through writing, we can do so to the very best of our ability. Put simply: we write to think better, and we think to write better.
This is not just a matter of professional responsibility. It is a matter personal enjoyment. Getting to think about subjects of consequence, getting to work through ideas and apply this thinking to the solving of real-world problems, should be a source of joy for the researcher, the teacher, the student. In pursuing the pleasure that comes with a life of the mind, we sharpen our thinking, making ourselves ever more into the kind of people society can trust, and be justified in trusting, with the development of the ideas that create and sustain progress.
In some ways, then, it would be easy to say that there is no role for AI in writing, thinking. But, at the same time, staying sharp, in accordance with our responsibility, means trying to understand anything which might help us develop our thinking, towards the goal of bringing more good ideas into the world (and making existing ideas better). And it is simply a fact that working with AI can yield much of genuine intellectual use. What if AI actually improves what we do? What if AI + human creativity deepens insight? It is the responsibility of the academic not to turn away from this, to keep exploring the use of AI, to keep seeing what it can do—what it can help us do—in the service of advancing human progress. AI, at its best, can function as a kind of intellectual whetstone. Through its prompts, through its capacity to gather data (which we should, of course, verify before using), through the assistance it can provide in helping us to say what we want to say better, it can help us do what writing should always do—aid us in the process of thinking, improve our ability to develop and express ideas and build a stronger intellectual foundation for all we do.
We can get carried away with this. Of course we can. These are heady days, an open road into the future, and AI may feel at times like a sports car we can exhilaratingly accelerate, leading to occasional skids. In the writing and thinking space, this can occur where norms have not yet fully cohered around AI use, creating confusion about what to do and what not to do. Is it OK to use AI to outline research writing and articles so long as this use involves deep intellectual engagement and shaping on our part? How far can we pursue this engagement, provided we are transparent about it? Academia has been wrestling with these questions, and academic public health is no exception. To my thinking, engaging transparently is indeed the best rule of thumb always, and particularly in this evolving landscape, as well as holding to the simple standard that we should not let AI do any thinking we should want to do ourselves. Writing being a form of thinking, this certainly applies to our use of words and our engagement with ideas. That is why including a “How I make this” statement for AI use on Substack, for the sake of reader awareness, seems to me a reasonable and practical approach to AI in the writing and thinking space.
But, as we well know by now, we are only beginning to understand the potential of AI. What this technology means for the world, for the development of the human being, is a book that has not yet been written. It is our task to meet this future with curiosity, hope, and a commitment to the integrity of our work and to doing all we can with the tools available to us, towards building a world where all can indeed thrive.
__ __ __
One concern I would add is that AI may change not only how we think, but what we value in thinking.
When answers come quickly, we may become less patient with uncertainty. When every idea can be polished instantly, we may become less tolerant of rough, unfinished thought. And when AI can make almost anyone sound articulate, it may become harder to tell the difference between clear writing and deep understanding.
There is also a risk of sameness. If millions of people are asking similar systems for help, could our writing slowly become more polished but less personal, less surprising, and less original?
So perhaps one of the challenges of the AI era is not simply protecting human thinking, but protecting the messiness, individuality, doubt, and struggle that often make good thinking possible.
The real test may not be whether AI helps us produce better work. It may be whether, over time, it helps us become better thinkers.
Thank you for this part of the series.