We humans always prefer to clean up the mess we create instead of avoiding it in the first place. Recycling, for example, seems like a more popular solution than consuming sustainable products. As appealing as these after-the-fact solutions are, they are always less effective because they mask or ignore the real problem.
So, it should come as no surprise that the next promise in communication and productivity takes the form of AI tools that promise to clean up the communication mess we have created with our own ten fingers. We are bombarded with emails, invited to too many meetings, and drowning in too many documents. What could be better than letting a machine handle most of this clutter and let us focus only on the important stuff? Assuming, of course, the machine “knows” what is important to us.
Every day we wake up with a new tool that promises to make communication easier using the power of Large Language Models like Chat-GPT. Got too many emails? No problem! We have a tool that will summarize them for you. Overloaded with meetings? Guess what! We have a machine that can extract the important stuff (and some action items), so why bother? Need to give a talk? Fear not! You just need to sketch a few slides, and magical software will write the speaker notes for you while you do something more meaningful. Work has never been so efficient. In the short run, all these new technologies will undoubtedly save us time. We will read less, write less, and probably interact less. What could be better?
The problem is that just like the superficial Meet-Less Monday solution to the meeting overload problem, we are trying to clean up the mess instead of fixing things from the ground up. All these “solutions” avoid dealing with the core of the problem: we don’t communicate effectively. Instead of trying to improve that, these technologies assume we will keep generating waste, and the only thing we can do is extract the important stuff as quickly as possible from the tons of redundant words we exchange. Nothing in these tools can make communication inherently better. The best we can hope for is that these technologies will help us identify the insipid parts of our interactions based on their “understanding,” ignore it, and never look back. Instead of making every word count, AI can help us find a few words to pay attention to in our messy interactions.
You might think this is a viable approach to make us more productive. But of course, AI tools will not be used only to read “better”; some are designed to help us “write” with less friction. It is not unlikely to assume that in parallel to the increased efficiency in reading, our inboxes will soon become even more cluttered with auto-generated texts simply because they cost nothing to generate. As often happens with technology, we might find ourselves in the exact same spot in terms of productivity and time wasted on meaningless communication. Instead of striving for better results, we are just building a machine that feeds itself with text, helps us process it, responds automatically, and so on.
But that’s not even the worst of it.
Shorter is Not Always Better
When I was in high school, I didn’t like reading much. We didn’t have even a fraction of the distractions we experience today, but I didn’t have the patience to read a 300-page book, let alone one chosen for me. So, when I had to prepare for an exam in Literature, I naturally used the one thing that is better than a book: a professionally-written excerpt. And it worked! To pass the test (and the occasional book report), you really didn’t have to read 300 pages. Ten percent of that was perfectly fine. It was the efficient thing to do.
But of course, there is more to literature than passing a test and more to books than a summary of their plot. A good story provides a rich and deep experience — much more profound than just the storyline and a flat description of the characters. A well-written text can strike a chord in the most unexpected places. A nuance that no one else might have noticed can open your eyes or touch your soul. When you settle with “the important stuff” and let someone else decide what’s important, you lose these opportunities. The shorter text was probably more efficient for the purpose of passing a test, but in terms of value, the unmediated original version, as written by the author, is irreplaceable.
Of course, an email is not a literary masterpiece, and a meeting is not an Academy Award-winning movie. Our inboxes are full of texts we can certainly live without, so why not look for the most efficient (and shortest) version of it? Why listen to every word in a meeting when we can have an automatically generated summary? Why even bother to attend it?
The reason, which most people seem to ignore as they are awestruck with the automation wonder, is that well-crafted emails and well-managed meetings are valuable in their original form if you believe that every word counts. I believe that if someone spends time writing something, there is value in reading it; if someone takes the stage in a meeting, there is a good reason to listen. I know you must be thinking, “In the emails I get, most words do not count!” And you are probably right. I’m not naive; I know our typical interaction today is ineffective because many of us don’t write and communicate well. But that is what we should work on: we must strive to make every word count. We must manage meetings so that every minute counts. There is more value in making our communication more meaningful than accepting our ineffective interactions as a fact of life and letting a piece of software decide what is essential for us. Automatically capturing a tedious meeting or a long email in five bullets is not a replacement for well-thought-out communication; it just helps us ignore the real problem.
Excerpts could have value. If you were present and engaged in a meeting or read the original email, a good summary could be used to recall some stuff later. The problem is that the benefit most people see in AI-generated excerpts is in skipping the original version altogether. Like the book excerpts I used in high school, no one will read the original email and then its summary. Instead of learning to write better emails and manage more effective meetings, we will keep generating communication waste knowing a machine will clean it up.
Reading and Writing are Essential to Thinking
Reading and writing are not just technical activities we can delegate to a machine — they are an essential part of our thinking process. Reading and writing are inseparable from our capacity to process and think things through. Something happens in our minds when we read and write. Even if much of what we read and listen to is redundant, deciding what to focus on is an act of processing. When we skip this part and let a machine do a significant portion of the processing for us, we diminish our capacity to think.
Did you ever participate in a meeting where something mumbled between slides struck a chord, and you found it far more revealing than what was formally on the agenda? Did you ever read an email where how something was described was more important than the facts and figures? Communication between humans is full of nuances. Sometimes they are under the surface, in the tone things are said or in what isn’t being articulated; in other cases, these nuances are just there on the table, but a “summarizing machine” will “consider” them to be nothing but noise. If we never read the full text or listen to what is being said, we lose our ability to identify and act upon these nuances. It is not rare that such allegedly insignificant fragments of the text can trigger innovative ideas and breakthroughs; they are opportunities we cannot afford to miss.
Letting AI write for us is the other side of the problem. On the surface, these advanced tools can articulate precisely what we aim for. All you have to do is craft the right prompt (or, in the case of the auto-generated speaker notes, craft the right slides). But the output of any machine that is out there lacks one crucial thing: what’s really on our mind. The text we write is a reflection of our thinking process. Hopefully, it is a well-crafted and organized version of it and not just a brain dump. AI can emulate something that looks like it, but it does not understand what is really on our minds, so it will always be a shallow representation of our ideas. We can certainly instruct it to generate texts in different styles and tones, but these are nothing but imitations of generic representations of how other people sound. AI will never capture the nuances of our thought process.
Now, the typical approach (at least at the time of writing this) is that AI could be used to generate a first draft, but you should edit it to make it more accurate and more personal. Let’s be honest, though. As these tools become better and more widespread, no one will do that. We will use the generated text almost as is, and even if we edit it, we will not think about it deeply. Not when our primary goal is to save time. Whatever words we will send out to the world, they will not be our own. We won’t be the ones writing them, and we won’t be the ones reading them. We will become nothing more than components in a communication machine that feeds itself.
The Human Factor
So AI can write emails for us and read the emails we receive and summarize them; AI can generate the speaker notes for a presentation and then summarize the meeting for the other participants. Notice something is missing? Under this vision, we humans are not really an essential part of the interaction; maybe we are not needed at all. The tools we’ve created will soon become the primary actors in our communication flows.
Communication is an overloaded term. We think of communication as the exchange of information, but this definition is derived from the machine world — the same world that created these amazing tools for us. Between humans, communication is much more than that: Communication involves listening, empathizing, understanding, and co-creating. We think in abstractions and metaphors, and much of our communication is based on these linguistic (and thinking) tools. When machines govern our communication and communication is focused solely on exchanging information, we lose the most important thing that brought us to this point: human interaction.
I find it funny that the same tech companies that in the post-pandemic era argue we should return to the office and break the digital barriers that until recently allowed us to work together remotely are now introducing the technology that is likely to eliminate the human factor from our day-to-day communication. I personally enjoyed remote work, but I never thought of letting Google, Meta, or Microsoft read and write for me. Sure, for the purpose of work, I meet fewer people physically now, and for the most part, I rely on Zoom, Teams, and a lot of text-based communication. But there are always people at both ends of the line. It is we who communicate. We talk, we listen, we read, we write. It is we who notice, process, and think.
I love technology and always look for ways to improve my work. But letting AI manage our interactions for us is not just a game-changer; it is taking humanity out of the equation.