Definition: A feedback cycle where AI tools observe misaligned behavior, interpret it as normal, and reinforce it—scaling drift invisibly over time.
Why it Matters: Drift becomes self-reinforcing. Your AI learns the wrong thing, teaches it back, and calls it improvement.
Common Mistake: Training AI on usage patterns without validating against aligned standards.
Related Concepts: AI Drift Scaling, Execution Blindness, Reinforcement Gap
Sample Usage: “Our AI learned from what they did—not what we wanted. That’s how the AI Drift Loop hijacked our execution model.”
Synonyms:
ai-drift-loop