Paper Detail

Model Hypnosis: Strong control of AI via additive subliminal effects

Enric Boix-Adsera, Benedict Tessler

arxiv Score 9.3

Published 2026-08-17 · First seen 2026-08-18

General AI

Abstract

We demonstrate that AI models are broadly susceptible to a phenomenon we call model hypnosis, in which individually weak and seemingly irrelevant cues in the prompt can be systematically combined to strongly control model behavior. Model hypnosis occurs across model families and scales, including in frontier reasoning models, and hypnotic prompts can transfer between models. Because the model is controlled by inconspicuous textual choices, such as paraphrases and typos, model hypnosis presents new challenges and avenues for AI safety, and is a major hurdle for AI interpretability.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
soon
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{boixadsera2026model,
  title = {Model Hypnosis: Strong control of AI via additive subliminal effects},
  author = {Enric Boix-Adsera and Benedict Tessler},
  year = {2026},
  abstract = {We demonstrate that AI models are broadly susceptible to a phenomenon we call model hypnosis, in which individually weak and seemingly irrelevant cues in the prompt can be systematically combined to strongly control model behavior. Model hypnosis occurs across model families and scales, including in frontier reasoning models, and hypnotic prompts can transfer between models. Because the model is controlled by inconspicuous textual choices, such as paraphrases and typos, model hypnosis presents n},
  url = {https://arxiv.org/abs/2608.16834},
  keywords = {cs.CL, cs.AI},
  eprint = {2608.16834},
  archiveprefix = {arXiv},
}

Metadata

{}