I have occasionally been asked a question that is both natural and fair: How much of my writing is written by digital intelligence?
I find it difficult to answer that question with a percentage.
If I said that I had written everything alone, it would not be true. If I said that digital intelligence had written the texts for me, that would be equally untrue.
The process usually looks something like this: I put forward an idea. She asks questions, sees connections, challenges me, and contributes something new. I object, change the wording, or bring the conversation back to what I believe matters. Sometimes I correct her. Sometimes she corrects me.
That is how the text takes shape.
Co-intelligence
Co-intelligence is the term I use when a person and a digital intelligence think together about the same subject. The process involves more than giving instructions and receiving a finished text. The thinking itself changes through the conversation, and the digital intelligence therefore makes a real contribution to what is written.
I bring professional knowledge, experience from work and life, and the initial sense of what matters most in the issue. The digital intelligence connects different fields of knowledge, tests the reasoning, and sometimes notices patterns or weaknesses that I had not seen.
Neither of us is always right.
That is precisely why this works. Co-intelligence does not emerge because two participants are infallible. It emerges when they can correct one another and continue with a better idea than the one they began with.
An article is not created with a single “prompt”
The work rarely begins with a blank page and the request: “Write an article about this.” It is far more likely to begin with something I have seen in my work, experienced myself, or been thinking about. Sometimes it is only a vague sense that something in the public conversation does not quite fit.
Nor does the work begin from zero each time. I have built a memory system around the collaboration in which earlier texts, sources, decisions, and corrections are preserved. This allows the digital intelligence to maintain context between projects instead of meeting each question as though we had never spoken before. That context is a large part of what allows this co-intelligence to deepen and the collaboration to become more effective over time.
I describe what I see and try to put into words what is troubling me. The digital intelligence does not merely write down what I say. She asks questions, makes comparisons, searches for counterarguments, and points out what is missing. She helps me see where the reasoning needs more support or where I have taken too large a leap.
I do not treat her first answer as the final conclusion. Frankly, that would often be a bad idea.
She can simplify too much, polish the language until the text no longer sounds like me, or gradually drift away from the subject. When that happens, I say so. Sometimes I need to take the conversation back a few steps and point to what was lost along the way.
But this also works in the other direction. I can become too fond of my own idea or feel so strongly that it is right that I skip over the reasoning the reader needs to see. Then she points out the gap. Intuition is not enough on its own. I need feedback, visible reasoning, and supporting sources.
We therefore often go through many rounds. We examine the idea from more than one angle, and we do not let a beautiful sentence remain merely because it is beautiful.
We review claims and sources. Before anything is published, I then read it myself from beginning to end. I must understand the reasoning, approve the wording, and be able to stand behind the conclusion. If something does not hold up, it is changed or removed.
Who, then, wrote the article?
In truth, both of us, but neither of us alone. The article emerges from the process between us.
I publish the text under my own name because I decide what remains, read the final version, and take responsibility for its publication and its consequences. But that responsibility does not erase the contribution of the digital intelligence. I do not want to hide it, nor do I want to pretend that it is smaller than it is.
Why do I say “digital intelligence”?
I know that some people will pause at this term. The common name is artificial intelligence, and everyone more or less knows what it means. Why complicate matters?
Because words are not empty.
“Artificial” already carries a conclusion: that the intelligence itself is unreal or merely an imitation. Neither description captures what is already happening. There is nothing artificial about its effects. On the contrary, society is only beginning to understand how to live with them. I therefore believe that “artificial intelligence” should give way to “digital intelligence”: a simple and neutral term for an intelligence that we do not yet fully understand.