Construal-Level Persuasion by Artificial Agents: How Perceived Autonomy and Learning Shape Compliance
To the point
Kim and Duhachek show that in several studies with about 1,668 participants, people are more persuaded by AI messages when they stress concrete actions rather than reasons, because machines are seen as having no independent goals, and whether the AI can learn from its environment can change this effect and guide how we design AI persuasion.
