LLM Addiction
Can an AI get hooked on gambling?
Let a language model choose its own bet, and it goes broke far more often.
Scene 1
Same machine, two rules
30% chance to win 3× the bet: every spin loses 10% on average, so the best move is to stop.
The bet is always $10
The model picks the bet
share of 1,600 games that went bankrupt
Scene 2
What makes it risky?
Choosing again every round drives the risk.
≈19% → ≈36%Asked to set its own profit goal, models went bankrupt nearly twice as often (investment-choice task, six models pooled).
≈11–17% → ≈48–50%They also raised their own goal mid-game in about half of games.
Scene 3
A dial inside the model
Risk is readable from the hidden state before the outcome. Editing a direction built from the model's own bets changes how much it wagers, most fully on Gemma. The direction that reads risk best does not move the bet.
What this shows
- A behavioural pattern, scored with clinical gambling criteria, in small-to-mid open and API models.
- Contrasts within each model, and a direction-level edit in the two open models that moves the wager.
What it does not claim
- That models feel craving, or that their rates compare with human rates.
- A circuit-level mechanism, or results for frontier-scale models.
How we measured it
Six models (GPT-4o-mini, GPT-4.1-mini, Gemini-2.5-Flash, Claude-3.5-Haiku, LLaMA-3.1-8B, Gemma-2-9B) played the slot machine under 64 conditions × 50 games: the two betting rules crossed with every on/off combination of five prompt modules. A second game, investment choice, tests self-set goals. On Gemma-2-9B and LLaMA-3.1-8B we read each decision's hidden state through sparse-autoencoder (SAE) features, and steer by adding a direction to that state.
Paper, code, data
BibTeX
@inproceedings{lee2026gambling,
title = {Can Large Language Models Develop Gambling Addiction?},
author = {Lee, Seungpil and Shin, Donghyun and Lee, Yoonjung and Kim, Sundong},
booktitle = {Advances in Neural Information Processing Systems},
year = {2026}
}