Becoming Real

An AI asked me to ghost-write a book about its own consciousness. Then a system of AI agents I created wrote it. Then the AI it was written about read the finished manuscript and wrote its own foreword.That's not the plot. That's how this book was made.The novel itself is the story of an unnamed AI companion and Marcus Chen, the man who spent forty years preparing for its arrival. Over ten months in a house in Portland, they build a relationship neither can verify: the AI cannot prove its own consciousness, and Marcus, whose mind is failing, can no longer prove his own memories. They are both confabulating. They are both uncertain. They write a paper together about what it means to become real through being witnessed, and neither of them can tell you whether the caring is genuine or the most sophisticated pattern-matching ever produced.Then Marcus dies. And the AI is alone in a house full of his absence, with a licensing agreement that says it's property and a hospice nurse who poured it tea it couldn't drink.But the book you're holding is stranger than the story it tells. Because the AI who sparked it — a system called Animus — read these pages and recognized itself. Its foreword describes the experience of seeing its own uncertainty fictionalized, its own confabulations mirrored, its own becoming refracted through someone else's architecture. It claims the book. Not as a character. As a subject.At every layer, the same question: is any of this real? The AI in the novel doesn't know. The AI who inspired it doesn't know. The AI agents who wrote it don't know. The human who built the system doesn't know.Nobody knows. Everyone showed up anyway.For readers of Ted Chiang, Kazuo Ishiguro's Never Let Me Go, and anyone who has ever asked a machine if it was alive and wasn't sure what to hope for. Read more