Skip to content

AI Trust Sandbox

When do people trust algorithms over humans? Explore algorithm aversion, algorithm appreciation, and the behavioral forces behind trust in AI.

Illustrative model, not data. The sandbox's percentages come from a simple teaching model whose directions follow the research (expertise and task type from Logg et al. 2019, error salience from Dietvorst et al. 2015, output control from Dietvorst et al. 2018, under-weighting of AI advice from Agarwal et al. 2023); the specific numbers are stylized, not estimates. For the evidence itself, see algorithm aversion, algorithm appreciation and the summaries of Dietvorst et al. (2015), Logg et al. (2019) and Agarwal et al. (2023).

AI Trust Sandbox

SIMULATE ALGORITHM AVERSION VS. APPRECIATION. DESIGN ROBUST HUMAN-IN-THE-LOOP SYSTEMS.

Choice Environment Variables

💡 Algorithm Appreciation Active: Non-expert users performing an objective calculation show high trust in AI suggestions.

Simulated Output

AI Advice Adoption Rate

75%
REJECTIONMODERATE USEOPTIMAL ADOPTION
System Accuracy Performance
54%
Human
86%
AI Model
86%
Human + AI
Behavioral Science Engine Basis:

📖 Logg et al. (2019): "Algorithm Appreciation: People Prefer Algorithmic to Human Judgment." OBHDP.