A few weeks ago, I wrote an article about the value of a liberal arts education in the age of AI. Citing the importance of critical thinking, analytical reasoning, persuasive writing, and collaborative problem solving, I argued that AI increases the premium companies will pay for human judgment, something that a liberal arts education cultivates.
The feedback that I received was positive. As a liberal arts graduate, that shouldn’t have been surprising. I have a few friends with engineering backgrounds, but my liberal arts colleagues outnumber them. Many of them were also positive about the six recently published articles that I cited.
I deliberately used recent sources to support my liberal arts argument. I didn’t have to. For example, for more than 40 years, MIT computer science professor (and AI pioneer) Patrick Winston delivered a lecture titled How to Speak. In his introductory remarks, Professor Winston stated that a person’s success in life will be determined largely by their ability to speak, their ability to write, and the quality of their ideas, in that order. As Professor Winston formulated and refined his lecture over the years, AI was available only to a select few.
From Knowledge Scarcity to Knowledge Abundance
For most of modern higher education, the system rewarded those who knew more. Content was scarce, and access to information was limited. Faculty served as gatekeepers of knowledge. Thanks to AI, that world is gone. Those of us who have used AI know that it can produce essays in seconds, passable research summaries, working computer code, and presentation outlines.
The constraint is no longer access to knowledge. It is the ability to make sense of it. Individuals who will stand out are not those who can generate the most content, but those who can clarify, interpret, and communicate it effectively.
The Empowerment Promise
Professor Winston introduced a practical concept for speakers called the empowerment promise. The promise is a tactic where the speaker tells the audience what they will be able to do or understand by the end of the talk. Interestingly, this is a discipline of thinking, not just speaking.
Artificial intelligence can generate answers to almost any question posed to it. However, it is not good at answering philosophical questions such as “Why does this matter?” Thankfully, humans can answer questions like that. Winston emphasized simple but powerful techniques in his advice about speaking. The key ones were:
- Clear structure
- Repetition for emphasis
- Signaling transitions
These tactics outline a stylistic approach to crafting a speech or lecture and can be powerful cognitive tools.
In the age of AI-generated content, structure becomes even more valuable. Machines can generate information quickly, but they do not consistently prioritize what matters most, nor do they organize ideas for maximum clarity. People who can turn complexity into coherence become indispensable. The role of the educated individual is shifting from generator of knowledge to interpreter of knowledge.
Attention Has Become the Scarcest Resource
Professor Winston was known for insisting that his audiences avoid distractions. He asked that his students close their laptops and turn off their cell phones. His reasoning was straightforward: people have a limited capacity to process language, and divided attention diminishes understanding. Not only that, he pointed out that the presence of laptops also distracted him as the speaker and reduced his effectiveness.
Winston’s observations have become more relevant. In fact, many high schools across the country have recently banned cell phones in classrooms to improve learning quality. Today, we are not just competing with other speakers or writers. We are competing with:
- Algorithmically curated feeds
- Constant notifications
- An endless stream of AI-generated content
In today’s environment, delivering information is not enough. Communicators must find ways to earn attention and hold it long enough to make an impact.
The AI Paradox: More Content, Less Clarity
Artificial intelligence has created a paradox. We generate more content than ever before. But comprehension does not scale at the same rate. As volume increases:
- Attention fragments
- Depth decreases
- Meaningful understanding becomes harder to achieve
As many people know, Large Language Models (LLMs) are the technology behind AI tools like ChatGPT, Claude, Gemini, etc. The models work by ingesting billions of pieces of content, then training the model to search for an answer to the prompt and write its response using probabilistic algorithms. But what if the content generated is written in an outdated style?
The Misunderstanding at the Heart of Writing
Winston’s lecture reminded me of a lecture from Larry McEnerney, the Director of the University of Chicago Writing Center, that has been circulating online for years. Professor McEnerney maintained that most of us were taught, explicitly or implicitly, that writing is about expressing what we know. As students (K-20), we were rewarded for things like demonstrating knowledge, organizing ideas, and writing clearly structured essays.
Those of us who followed instructions well received good grades and achieved academic success. However, Dr. McEnerney posited in his lecture that there is a problem with this style of writing.
The traditional model of writing that most of us were taught only works in an environment where someone is required to read what you write.
In school, your teacher or your professor was required to read your paper. In the real world, no one is required to read your paper.
McEnerney made a powerful claim. Writing is not about expressing your ideas. It is about creating value for a reader. If your writing does not solve a problem, improve understanding, or help someone make a decision, then it has no value to the reader. If your writing has no value, it will not be read.
Smart People Often Struggle
McEnerney was direct in his observation that highly educated people are often the worst writers outside their domain. This isn’t because they lack intelligence, but because they were trained in the wrong model. Highly educated people learned to write for instructors, critical evaluation, and practical demonstration.
When these highly educated people inevitably encounter the real world, they often find themselves writing for starkly different and more diverse audiences, like busy professionals, decision-makers, and users who have no obligation to read anything.
McEnerney stated that this is not a gap between writing models but is a structural flaw in how colleges prepare students for the workforce.
Effective Writers Operate From a Different Starting Point
Effective writers do not ask “What do I want to say?” They ask, “What problem does my reader need solved?” This question leads to a totally different writing structure.
Traditional academic writing demands rigor, but structurally, it can be perfunctory. We expect background on the topic, detailed research of the relevant literature, and a polished thesis.
In contrast, real-world writing that’s bogged down by research and “inside baseball” language, you’re likely to lose attention quickly. Effective real-world writing demands problem-solving, an understanding of the stakes involved, and an explanation of how your ideas lead to a solution.
The shift from topic-centered writing to problem-centered writing is the difference between being read and being ignored.
Why This Matters More in the Age of AI
Until recently, poor writing could still find an audience. That is no longer the case. We live in a world where millions of pieces of content are generated every year. Last week, I reported that more than four million books were published in the United States in 2025. The marginal cost of producing “acceptable” writing is approaching zero. Artificial intelligence can already:
- Summarize information
- Organize ideas
- Produce grammatically correct prose
But AI struggles to:
- Identify a meaningful problem for a specific audience
- Establish urgency or stakes
- Deliver insight that changes how a reader thinks
In other words, AI can replicate the academic writing model remarkably well. AI struggles with the value-creation model. The primary reason for this is that Large Language Models (LLMs) create their probability-derived responses based on billions of ingested articles and books. Naturally, most of those articles and books were written in the old style.
The Emerging Divide
This creates a new divide, one that I believe will become increasingly visible in the workforce. There are two sides. On one side are individuals who can produce traditional content. On the other side are individuals who can create value through writing.
The first group will be augmented and, in many cases, replaced by AI because AI is trained to write like that group. The second group will become more valuable.
A Higher Education Problem We Can No Longer Ignore
From my perspective, this has direct implications for higher education. For decades, higher education has focused on teaching students what to know and evaluating their ability to explain it.
We have spent far less time teaching students how to communicate in environments where attention is scarce, how to write for audiences who are not obligated to listen, and how to create value through communication. More than a writing issue, this is a looming workforce readiness issue.
A Different Model for Teaching Writing
If we were to redesign writing instruction for today’s world, it might look like this:
- Start with the reader, not the writer
- Define a real problem that matters to that reader
- Make the stakes explicit
- Deliver insight that improves the reader’s position
- Measure success by impact, not completion
This is not how most writing is taught today. As McEnerney pointed out, it is how most effective writing works in practice. Ironically, it is also not how most LLMs generate their articles about various topics.
A Personal Reflection
Over the course of my career as a healthcare executive, a university president, and now a board member and investor, I have read thousands of reports, proposals, articles, and books. The ones that stood out were not the most polished. They were the ones that:
- Framed a problem clearly
- Made me see something differently
- Helped me make a better decision
That is what effective writing does. Effective writing is becoming increasingly important in the age of AI. Given the vast proliferation of content that no individual can possibly keep up with, effectively written content will be more valued by humans who serve as expert moderators of knowledge.
Parting Advice
Patrick Winston’s lecture, which I used to lead off this article, was primarily about speaking. However, in his introductory remarks for his How to Speak lecture, he noted that a person’s success will be defined by their ability to speak, their ability to write, and the quality of their ideas in that order.
Professor Winston’s Empowerment Promise, which advises the speaker to tell the audience what they will be able to do or understand by the end of the talk, is eerily similar to Professor McEnerney’s maxim that if your writing does not solve a problem, improve understanding, or help someone make a decision, then it has no value to the reader.
Winston was a Computer Scientist at MIT who directed the MIT Artificial Intelligence Laboratory from 1972 to 1997. While LLMs hadn’t been invented at that point, it’s likely that Winston saw the wave of content and knowledge generation coming and wanted to emphasize the importance of individuals mastering human skills like effective speaking and writing.
Larry McEnerney earned a PhD in English but didn’t want to spend his career writing about literature. He joined the University of Chicago’s writing program and, five years later, became its director, a position he held for nearly three decades. While McEnerney had no background in computer science or AI, his focus on effective writing is excellent advice for anyone using and adapting AI-generated content in the workplace.
We are entering a period where the ability to generate content and ideas is no longer scarce. The ability to create value through speaking and through writing is. Having the critical thinking skills to evaluate content, especially AI-generated content, and to communicate about that content to your bosses, colleagues, and team members will be crucial to your career success. Education leaders should make sure that their students learn how to evaluate AI content and output, edit, and communicate the ideas generated effectively. McEnerney advocates for effective writing. Winston advocates for telling the audience what they will understand at the end of your lecture or speech. Both tactics lead to value, and that’s what matters the most if you are working to establish your brand, your authority as an expert, and your professional career.
