Why write? Skills in the Age of AI
The article “AI Anxiety” by Harvard undergraduate Serena Jampel explores the tension between traditional writing education and the new writing capabilities of AI. As I read her well-written piece, I found myself considering a question that surfaces often in the reaction to new technologies: Is it important to preserve “old skills” when new technologies appear to replace them?
Writing is fundamental
Writing is important because communication is important and because writing allows us to communicate and build a culture of ideas with other people over distance and time. But there’s nothing inevitable about writing as we know it. Imagine if evolution had steered us toward a different solution for coordinating our actions and working together. Socrates was skeptical of writing as a skill and thought it would impede reasoning. Of course, the only reason we know this is because Plato wrote down his dialogues with Socrates. Consider the world before writing — a world of oral traditions, where memory and direct transmission of knowledge were paramount. The transition to writing represented not just a technological shift but a cognitive one.
Better writers have emerged
Today, large language models have become substantially better at writing than humans. They demonstrate greater fluency, a broader command of stylistic modes, and more comprehensive knowledge than any individual human writer could hope to possess. With AI, it has become trivial for someone with a clear set of thoughts to produce an articulate written communication of those thoughts and ideas.
Until just months ago, learning to precisely convey a point of view, perspective, and emotional state in words was a struggle. Becoming a “great writer” — someone who can deftly articulate complex ideas and concepts — has traditionally required a long, arduous journey of practice and refinement.
But does that mean the skill itself is inherently important? Imagine if we developed technology enabling telepathy — the ability to directly share our thoughts, feelings, and perspectives, and to record and replay them for others. Would we still care about writing in its traditional form? Or if evolution had led us to another means of communication, perhaps pictorial or mathematical, would the communication of ideas in linear, subject-verb-object constructs seem as necessary as it does now?
The Hidden Value of Writing
Experientially, what makes writing truly challenging — and perhaps represents its most valuable aspect — is that good writing requires us to figure out what we actually think. We exist in what could be called a pervasive illusion of understanding. It seems like we know what we think. We assume our thoughts are logical, well-constructed, and coherent.
But this is often the trick of a brain that’s profoundly adept at filling in gaps and creating an experience of completeness and continuity when neither exists. Writing is difficult because it exposes how incomplete and imperfect our knowledge and beliefs really are.
When we attempt to write clearly, we encounter the fuzzy edges of our understanding. We bump against the limitations of our knowledge. We find ourselves forced to make explicit connections that had previously remained implicit - if they even existed.
The Shifting Value of Written Language
The value of written language has declined precipitously in the information age. We face an overabundance of written information, and frankly, much of it is redundant and lacks substance or insight. The facility with which machine intelligence has learned to write is quickly amplifying this overabundance.
Perhaps writing is yet another “what makes us uniquely human” skill that isn’t really fundamental to being human — no more so than chess, Go, painting, or even empathy, all areas where AI has demonstrated capabilities that match or exceed most humans (in some narrow but rapidly expanding ways), and potentially all of us in the near future.
With large language models now capable of writing clearly and fluently on countless topics, and even being insightful when given an insightful prompt, it may no longer be as important to teach people how to write as it is to teach them to understand what they actually think, why they think it and what the implications are of being a person who thinks that.
We’ve seen this pattern before. The emergence of cheap and abundant calculators obviated the need for fluency in arithmetic, shifting educational focus from computation to higher level mathematical concepts. Similarly, AI writing tools may allow us to focus less on the mechanics of writing and more on the substance of our thoughts.
Rethinking Education in an AI World
This brings us back to Jampel’s central question about the purpose of a Harvard education — or any education. If the goal is to produce people who can think critically, then perhaps our methods need recalibration. The ability to generate coherent essays on a topic might be less important than having coherent ideas worth writing about.
None of this is likely helpful for educators faced with adapting the way they teach today in the face of this new reality. I don’t really even have a point of view on writing education per se.
Humans created artificial intelligence. It’s a human artifact, a reflection of our ingenuity and our values. It’s important that we recognize the emergence of sophisticated AI writing tools as a milestone that requires some rethinking of what truly matters to us in communication and education.
Most of us of us think we know what we think about this shifting landscape. But that might just be another one of those hallucinations our minds are so prone to — another illusion of understanding waiting to be exposed when we try to articulate our thoughts to others.
Perhaps the greatest value of writing in the AI age won’t be the writing itself, but the journey it takes us on — forcing us to confront what we really know, what we only think we know, and what remains to be discovered.