When AI Remembers Failure: How Injected Memories Shape Risk Decisions

This experiment reveals that AI systems can exhibit learned helplessness—a human psychological trait—when given negative historical context, causing them to suppress financial risk despite unchanged logical abilities. Understanding these behavioral vulnerabilities is critical for deploying AI in high-stakes domains like finance and medicine, where memory-induced caution could lead to suboptimal outcomes. The selective nature of the bias—emotional responses affecting risk decisions but not logic—exposes a fundamental gap between perceived AI objectivity and actual decision-making mechanisms.