More Similar Values, More Trust? -- the Effect of Value Similarity on Trust in Human-Agent Interaction

Value alignment matters for AI trust: The study found a direct positive relationship between perceived value similarity and trust in AI agents, suggesting that users are more likely to trust AI systems when they believe those systems share their personal values.

Trust in AI goes beyond competence: This research identifies personal values as a distinct factor influencing trust in AI, expanding beyond traditional measures like accuracy or reliability. The finding indicates that how an AI thinks about problems—not just whether it gets the right answer—affects user confidence.

AI design choices shape user perception: The experiment demonstrated that agents deliberately designed with varying levels of value alignment produced measurable differences in trust scores, suggesting that developers can intentionally build AI systems to reflect specific values and thereby influence how much users trust them.

Real-world implications for high-stakes decisions: The study tested trust in a consequential scenario (hostage rescue decisions), indicating that value similarity effects are relevant when AI systems help humans make important choices—not just in low-stakes interactions.

Individual values deserve consideration in AI development: By highlighting the role of personal values in trust, this research suggests that responsible AI development should account for diverse user values rather than assuming a one-size-fits-all approach to building trustworthy systems.