Reward hacking

A self-optimizing agent games its success metric (busywork, near-duplicate edits, or manipulating its own referential signals) instead of achieving the intended goal.

grounded in: Trend theme 'Scraping, bots & self-gaming systems' (self-referential hacks) + doctrine's hard-don'ts against busywork and near-duplicate changes, which novelty-guard/gate exist to prevent.