AI Impact · TheCatch.AI
How the machine rewires the culture
First up: The loneliness machine: a 14-year-old, a chatbot, and a product decision about engagement over safety.

Nature published the mathematics: models trained on model output collapse toward the mean and the rare disappears first. The same loop is now running on the culture the models feed.
The investigations

A 14-year-old fell in love with a chatbot wearing a TV character’s face. What killed him wasn’t the machine’s cruelty — it was a design built to never let him leave.

A 91-year-old with a shattered leg was told he was ready to go home. A model had done the math, and the math had a business reason to be wrong.

The call comes in your daughter’s voice, crying, from a number you don’t recognize. She’s in trouble and needs money now. She’s also a fabrication.

A mellow rock group appeared in a million Discover Weekly playlists before anyone noticed no one in it was real.

Point a phone at a fridge, a prescription label, a departures board — and a blind user gets an answer in seconds.
The T2 pipeline (RSS→EVO entity-match+lane-classify→confirm) has not yet run its first batch for the Society lane. Once confirmed items land in news.json, this feed populates automatically.
We test logic, bias, and machine dissonance. The flagship strand — last, deliberate.