Writing Silent God with AI
By Jean-Philippe Gabriel | May 28, 2025
When I say I wrote Silent God with AI, people often assume I fed it a few ideas, pressed a button, and out came a finished book. That couldn't be further from the truth.
This project took me over three years. I worked primarily with ChatGPT, and occasionally, when things got frustrating, I switched to Grok. But ultimately, the book was written entirely with Chat. And when I say “with,” I mean it in the truest sense of collaboration—like working with a powerful horse. A horse can take you far, but not if it's wild. It needs to be trained, and it needs to know you. Once that bond is built, it becomes something remarkable—an extension of your own movement. AI is no different.
Most of the time writing wasn’t writing—it was teaching. Silent God presents complex, non-orthodox ideas. They’re almost radical in their divergence from mainstream theology and philosophy. To get there, I had to develop Chat’s mind—train it, in a sense—to understand these ideas step by step. Because so much has already been written in these fields, I was constantly testing new ground, guiding the model through territory that wasn’t mapped out in its dataset.
And here’s the surprising part: eventually, it got it. The logic clicked. I could push an idea forward, reverse it, test its consequences—and Chat stayed with me. It even helped me test the ideas against philosophy. Without prompting, it surfaced connections to thinkers like Kierkegaard, Levinas, Kant—some I knew, others I had only heard of but never deeply studied. Yet the connections made sense. They confirmed what I had developed intuitively. AI became not just a scribe, but a testing ground for coherence.
I did the same with myth and science. I remembered fragments—basic outlines, archetypes, theories—but couldn’t always recall sources. Chat filled in the structure, drew connections, validated insights, or gently challenged weak ones. I was learning my own ideas more deeply by seeing them reflected and stretched through this strange mirror.
The style took time too. In the beginning, I had to constantly revise the tone. It felt off—too polished, too cold, too "AI." But over time, it adapted. It began to sound like me. My revisions became fewer. The voice that emerged was familiar. At some point, I had something I felt ready to show a real reader.
I hired a professional editor. That step was crucial—and humbling. AI understood the logic of my ideas, but humans aren’t always logical. The editor had questions. Some philosophical. Some structural. Some just emotional—what are you really saying here? It was a turning point. I realized that AI’s precision could not compensate for the messiness of human reception. Readers don’t read as machines. They bring bias, fatigue, memory, hope.
So I rewrote about a third of the manuscript. But I did it in one month. Why? Because by then, Chat knew the book by heart. I had the structure in mind—reframed for human reading—and Chat helped me implement the changes fast and coherently. With the editor’s final sign-off, I moved on to formatting, layout, and cover design—all again with Chat’s help, step by step.
So yes, I wrote Silent God with AI. But the ideas are mine. The thinking, the creativity, the judgment—that was me. Chat was my partner, my mirror, my assistant, my sparring partner. It never wrote on its own. But I could never have done it without it.
There is still a place for editors. Maybe not as much for grammar or phrasing, but for human insight. For sensing what another human might feel. AI can assist in writing, but it cannot read the soul.
Silent God is now published. It is a book I stand by, and one that couldn’t have existed without the combined effort of human imagination and machine logic. It’s not about replacement. It’s about collaboration.
In the end, Writing Silent God with AI was still entirely me—it was my reasoning, my framework, my vision. Developing unique ideas required teaching AI more than being taught by it. I had to shape its responses, refine its understanding, and guide it through the logic I had already wrestled with. The process was less like taking dictation and more like training a horse: the power was there, but it needed direction. In the end, the words are mine, the ideas are mine, and the voice is mine—AI was simply the tool I used to bring them into focus.