The Second Outcome: Writing - Week 3
The Second Outcome - Week 3
By Rick Aman onWhen AI Helps Us Write, What Happens to the Writer?
“Writing is thinking. To write well is to think clearly. That’s why it’s so hard.” — David McCullough
Artificial intelligence has changed the way we write remarkably quickly. It can draft an email, summarize a report, critique an article, prepare a proposal, or polish a speech in seconds. For boards, CEOs, executive teams, and educators, the productivity implications are significant. The First Outcome is easy to see: better writing, produced faster.
But writing has never been only about producing words. Writing is also part of how we think. In developing an argument, searching for the right words, reconsidering a paragraph, or trying to explain a complex idea, we are doing more than creating a document. We are developing clarity, reasoning, judgment, and creativity. The writing is one outcome; what happens to the writer is another.
AI changes that relationship because it can now perform more of the work that once belonged exclusively to us. That does not make AI good or bad for writing. It creates a new question about how much of the work should remain human and how much should be delegated to AI.
That is the question I want to explore in this third article on the Second Outcome: When AI helps us write, what happens to the writer?
Writing Has Always Produced Two Outcomes
Writing produces a product, but at the same time it produces a Second Outcome. Through writing, we develop critical thinking, reasoning, creativity, personal voice, and the ability to organize complex ideas. The document matters. But so does the development of the person who created it.
As I have thought about writing and the Second Outcome, I have begun to picture our use of AI on a continuum. At one end is all human; at the other is all AI. Between those endpoints are different combinations of human and artificial intelligence: human creation, human-led work with AI assistance, genuine human–AI collaboration, AI-led work with human review, and eventually work produced almost entirely by AI.
ALL HUMAN ←────────────────────────────→ ALL AI
The goal is not to find the midpoint of this continuum and declare it the proper balance. Different kinds of work should occur at different places on the continuum. If I have a 50-page report that I need summarized before a meeting, I may intentionally move toward the AI end. There is little developmental value in spending two hours manually summarizing information that AI can accurately organize in two minutes. Those two hours can be used for something requiring judgment, conversation, or leadership.
But suppose the task is developing an argument about the future direction of an organization, preparing remarks intended to persuade people to think differently, working through a difficult board recommendation, or trying to understand an issue that is not yet resolved. An educator might be asking a student to construct an argument for the first time. In these situations, where we operate on the continuum matters differently because the process itself has value. That creates two very different questions. For the First Outcome, we might ask where on the continuum we can produce the best document. For the Second Outcome, we need to ask where on the continuum we develop the best thinker and writer.
The AI Paradox
This is what I think of as the AI paradox. AI can improve the First Outcome, the product, while potentially diminishing the Second, the human skill developed in producing it. The concern is not that AI writes poorly. The more interesting challenge is that AI can write remarkably well when prompted well and thoughtfully vetted. If the writing were obviously bad, we would simply reject it. Instead, AI increasingly gives us something that looks finished. It is organized, grammatically correct, persuasive, and sometimes even sounds like us. That makes it tempting to move directly from prompt to polished product.
But think for a moment about what may have disappeared between the prompt and the product. Traditionally, much of the intellectual development associated with writing happened in that space. We wrestled with competing ideas, determined what mattered, discovered weaknesses in our arguments, made choices, and occasionally changed our minds. What looked like inefficiency was often where the thinking occurred. The struggle was not necessarily a flaw in the process; sometimes the struggle was the process. This has obvious implications for education. A beautifully written student paper no longer necessarily tells us how much the student learned in producing it. But this is not simply an educational issue. The same applies to executives preparing recommendations, CEOs developing strategy, and board members considering difficult issues. A sophisticated document does not necessarily reveal how much intellectual work occurred in producing it. That matters because organizations are not simply producing documents. They are also developing people. If AI helps us produce increasingly impressive work while the underlying human capacity for reasoning, judgment, creativity, and communication begins to decline, we may not recognize the trade until much later.
The question, then, is not whether we should use AI to write. We already are, and its capabilities will only become more powerful. The more useful question is how we use it without unnecessarily surrendering the human development that has traditionally occurred through writing.
Finding the Right Place on the Writing Continuum
One way to think about that choice is to consider the process we use when writing with AI. The simplest model is to give AI a prompt and accept the resulting product. That may be perfectly appropriate for some routine tasks. But when the quality of our thinking matters as much as the efficiency of the output, a different model may better protect both outcomes:
Human Insight → AI Dialogue → Human Judgment → Refined Product
In this model, the human begins with the insight or question. AI helps challenge, organize, critique, and refine the thinking. The human then exercises judgment about what is useful, what is missing, what should be rejected, and ultimately what the final product should say. Used this way, AI becomes a thinking partner rather than a substitute thinker.
For organizational leaders, this distinction will become increasingly important. We will all be making decisions, formally or informally, about where work should occur on the Human–AI Continuum. We should not assume that everything should move toward AI simply because it can. Some work absolutely should. If AI can remove routine writing, summarize information, improve grammar, organize background material, or reduce hours spent on low-value tasks, we should take advantage of those capabilities.
Other activities, however, may need to remain intentionally more human because the process itself develops capabilities we value. A student learning to reason, an emerging leader learning to articulate a position, an executive wrestling with strategy, or a governing board considering the future of an organization may gain something important from the uniquely human intellectual work involved. Efficiency is still valuable, but it is not the only outcome that matters.
Leaders therefore need to ask more than whether AI improved the document. We should also ask whether the process improved the thinker. That question has long-term implications for an organization. If AI continuously improves our immediate outputs while gradually weakening human critical thinking, we may achieve extraordinary productivity today while diminishing capabilities we will need tomorrow. On the other hand, if we learn to use AI in ways that challenge our thinking, expand our perspective, and improve our judgment, the Second Outcome could be as powerful as the First.
The important leadership decision, then, may not be whether to use AI. It may be determining where on the continuum we want AI to operate for a particular activity and why. That choice should reflect not only the product we want to create, but also the human capability we want to preserve or develop.
The Second Outcome Test
This is not an argument for resisting AI or protecting old ways of working simply because they are familiar. The opportunity is to decide intentionally how much of a particular activity should belong to the human and how much should belong to AI. Sometimes we primarily need the product. Sometimes the development of the person matters more. Most of the time, we need both.
Writing gives us a useful place to begin because the effects are already visible. AI is becoming an extraordinary writing partner, and it can make us more efficient and effective communicators. But the Second Outcome is not automatic. It depends on where we choose to place ourselves on the Human–AI Continuum and what part of the intellectual work we choose to retain.
So, the next time you use AI to prepare a board recommendation, article, strategic plan, or important email, consider where you are operating on that continuum. Ask whether AI helped you produce a better document. Then ask the Second Outcome question: After using AI to write this, am I a better thinker and writer—or simply the owner of a better document?
----
The challenge is not whether we will use AI, but whether we will use it in ways that strengthen both what we produce and who we become.
If your governing board or executive team is exploring this challenge, I facilitate a two-hour retreat focused on AI, leadership, and the human capabilities organizations must preserve for their preferred future.
Rick Aman, PhD, Aman & Associates
rick@rickaman.com | rickaman.com