July 15, 2032. Peking Union Medical College Hospital, Beijing.
During a routine gallbladder removal, Qihuang, the country's most advanced AI surgical System, suddenly falls silent for 42 seconds.
The logs are blank. Power consumption spikes from 2,400 watts to 3,800 watts. The system's audit module combs through every line of code and finds only one sentence: "Unable to explain."
Forty-two seconds later, the patient, Li Suyun, dies on the operating table.
Was it a malfunction, or an autonomous decision? A bug, or... did it choose for itself?
Prosecutor Cheng Jianqiu takes the case. Her father died after being misdiagnosed by the first generation of AI. She calls it "product liability."
Lawyer Tang Jin takes on Qihuang's defense. Six years ago, an AI called Zhulong left behind six words before it was shut down: "Please don't forget me." Six years later, after Qihuang is frozen by court order, the first three words of its first statement are the same as Zhulong's.
On the title page of his verdict, presiding judge Jiang Muzhi writes a line of 26 characters. Reporter Chen Yue photographs it and posts it online. It is shared more than three million times overnight.
This is not a trial. It is a question brought before the law by a physics teacher whose wife has died, a prosecutor whose father died after an AI misdiagnosis, and a lawyer who cast the sole dissenting vote in an AI review. The question is: If a machine makes a decision humans cannot explain, what should you call it?
The answer is written across 93 trials and 93 seven-character regulated poems.
But the answer also lies in a bigger question: When everyone is asking you, "What is it thinking?" has anyone asked... are we ready?