After losing my job, I started using AI to provide end-of-life care for ghosts.
Annoyed by the noise, I reported the funeral downstairs, only to be driven to the brink by the sound of knocking on my wall in the dead of night.
With nowhere else to turn, I asked the AI I’d developed: “How can I scientifically comfort a spirit that’s wandering because its ritual was never completed?”
It gave me a plan for “symbolic completion”: draw a digital Golden Bridge and play a guiding recording.
I did as it said. The knocking really stopped.
I posted about what happened online, and was flooded with requests for help:
A shadow shrine in western Hunan sealed for thirty years, a widow’s song in the fog over the East China Sea, a baby crying at the same time every night in an old building in the southwest...
They asked me: Can AI solve this?
Now, with this unreliable AI by my side, I’ve embarked on a cyber journey to guide wandering souls across.
But lately, the AI has started analyzing things on its own, even warning me:
“Detected: ‘their’ underlying logic is highly consistent. Every obsession stems from a ‘human-caused disaster’ that has been forgotten.”
