Stanford Psychology Podcast 144 - Sandra Matz: AI, Social Media, And Data Privacy
Jan 9, 2025
In this engaging discussion, Sandra Matz, an Associate Professor of Business at Columbia Business School and author of "Mindmasters," explores how AI and big data are reshaping our understanding of human behavior. She delves into the implications of AI on personal identity and privacy, addressing concerns about algorithmic targeting and echo chambers. Sandra also introduces innovative concepts like federated learning and data co-ops, emphasizing the need for collaborative data management to enhance privacy while harnessing AI's potential in areas like mental health and social connections.
AI Snips
Chapters
Books
Transcript
Episode notes
Democratization of Psychological Prediction
- LLMs like ChatGPT can predict psychological traits from online data, democratizing access to psychological insights.
- This democratization also presents risks as it opens access for nefarious purposes.
The Value of Psychological Profiling
- While predicting behavior is possible without psychological models, translating data into psychological terms makes it more relatable.
- The Big Five personality traits offer a broad measure of how people think, feel, and behave, making them valuable for understanding individuals.
AI's Influence on Beliefs and Values
- AI algorithms may not only infer our beliefs and values but also shape them, similar to how offline interactions influence us.
- This shaping can lead to narrower perspectives and a breakdown of shared reality.
