Stuck on Suggestions: Automation Bias, the Anchoring Effect, and the Factors That Shape Them in Computational Pathology

Abstract

Artificial intelligence (AI)-driven decision support systems can improve diagnostic accuracy and efficiency in computational pathology. However, collaboration between human experts and AI may introduce cognitive biases such as automation and anchoring bias, where users adopt system predictions blindly or are disproportionately influenced by AI advice, even when inaccurate. These effects may be amplified under time pressure, common in routine pathology, or shaped by individual user characteristics. We conducted an online experiment in which pathology experts (n = 28) estimated tumor cell percentages: once independently and once with AI support. A subset of estimations in each condition was performed under time strain. Overall, AI assistance improved diagnostic performance but introduced a 7% automation bias rate, defined as accepted negative consultations where previously correct independent judgments were overturned by incorrect AI advice. While time pressure did not increase the frequency of automation bias, it appeared to intensify its severity, reflected in stronger performance declines associated with increased AI reliance under cognitive load. A linear mixed-effects model (LMM) simulating weighted averaging showed a statistically significant positive coefficient for AI advice, indicating moderate anchoring on system output. This effect increased under time pressure, suggesting anchoring bias becomes more pronounced when cognitive resources are limited. A second LMM assessing automation reliance, a proxy for automation and anchoring bias, showed that professional experience and self-efficacy were associated with lower dependence on AI, whereas higher confidence during AI-assisted decisions was tied to increased AI reliance. These findings highlight the dual nature of AI integration in clinical workflows: improving performance while introducing risks of bias-driven diagnostic errors.

Figures (3)

• Figure 1: The study interface presented to participants during tcp estimation. Based on the experimental treatment, the AI component (comprising the AI prediction, model reasoning explanations through prototypes, as well as cell detection visualizations) and the countdown timer element were displayed. The arrow indicates how the user interface changed as the timer progressed, generating a sense of urgency and limiting user interaction once the countdown expired.
• Figure 2: Scatter plots illustrating how participants' independent assessments and AI recommendations, respectively, influence the final AI-aided tcp estimates. The slopes correspond to the fixed-effect coefficients from the linear mixed-effects model analysis (Model 1), with the sign and magnitude of the coefficient for the system prediction indicating the direction and strength of the anchoring effect.
• Figure 3: Bar plot illustrating differences in linear mixed-effects model coefficients for baseline estimates and AI predictions, comparing AI-assisted assessments conducted with time pressure (TP) (Model 1.2) and without (Model 1.1). The coefficient for AI advice reflects the degree of anchoring on system output.