Build Trustworthy AI: Measure Human Connection

סקירה כללית

Transform how you approach AI companion products by mastering the psychology of human-AI trust. You'll apply anthropomorphism concepts to categorize the human qualities users attribute to AI—classifying signals as warmth-related or competence-related to build actionable user profiles.

You'll master the two-question warmth-and-competence survey that captures how users perceive your AI, rooted in decades of social psychology research. And you'll apply threshold-based decision rules that transform passive data collection into proactive user protection.

Through realistic Role Plays with skeptical colleagues and program coordinators, you'll practice categorizing anthropomorphic signals and presenting data-driven recommendations. Coach Dialogues challenge your judgment on ambiguous cases where the data doesn't tell the whole story.

Essential skills for UX researchers, Trust & Safety analysts, product managers, and anyone building or supporting AI companion experiences. No technical background required—just a commitment to responsible AI deployment.

סילבוס

• Apply Anthropomorphism Concepts and Measure Trust
Apply the psychology of anthropomorphism to real AI companion interactions—categorizing the human-like qualities users attribute to chatbots and classifying each signal as warmth-related or competence-related. You'll use the Anthropomorphism Identification Worksheet to document findings from chat transcripts and connect those signals to the survey instrument you'll administer.
• Apply Surveys and Evaluate for Human Escalation
Master the complete workflow from administering the warmth-competence survey to making threshold-based escalation decisions. Learn why an average score below 4 signals the need for human outreach, and how to flag accounts appropriately in your Dashboard. You'll practice the full cycle: send survey, record results, evaluate against threshold, and document escalation decisions.

---

נלמד על ידי

Coursera Support and Ritesh Vajariya

---

נושאים

Artificial Intelligence