.svg)
What is a Feedback Loop in Machine Learning?
A feedback loop in ML arises when a model’s predictions change user behavior or data collection, which then retrains or influences the next model—potentially amplifying bias, gaming, or error.
Recommendation and content systems are classic examples; generative assistants can create sycophancy loops.
Related Giskard articles
Detect harmful feedback dynamics with Giskard — probe assistants for sycophancy and measure quality drift over time. See Sycophancy in LLMs · giskard.ai.
