We study a stylized social learning dynamics where self-interested agents collectively follow a simple multi-armed bandit protocol. Each agent controls an ``episode": a short sequence of consecutive decisions. Motivating applications include users repeatedly interacting with an AI, or repeatedly shopping at a marketplace. While agents are incentivized to explore within their respective episodes, we show that the aggregate exploration fails: e.g., its Bayesian regret grows linearly over time. In fact, such failure is a (very) typical case, not just a worst-case scenario. This conclusion persists even if an agent's per-episode utility is some fixed function of the per-round outcomes: e.g., min or max, not just the sum. Thus, externally driven exploration is needed even when some amount of exploration happens organically.
@article{arxiv.2602.05835,
title = {Bandit Social Learning with Exploration Episodes},
author = {Kiarash Banihashem and Natalie Collina and Aleksandrs Slivkins},
journal= {arXiv preprint arXiv:2602.05835},
year = {2026}
}