Sandra Matz

AI Summary

→ WHAT IT COVERS Computational social scientist Sandra Matz explains how companies use digital footprints—social media posts, credit card data, location tracking—to build psychological profiles that predict personality traits and influence purchasing decisions, voting behavior, and personal choices through targeted messaging. → KEY INSIGHTS - **AI-Powered Profiling:** ChatGPT can analyze public social media posts to generate Big Five personality profiles (openness, conscientiousness, extroversion, agreeableness, neuroticism) without specialized training data, democratizing psychological targeting capabilities that previously required datasets linking questionnaire responses to digital traces. - **Personalization Effectiveness:** Matching marketing messages to personality traits increases conversion rates by approximately 50 percent, as demonstrated in beauty retail experiments where extroverts responded to dance-themed ads while introverts preferred solitary messaging, with similar results achieved in financial savings programs targeting low-income households. - **Data Collection Sources:** Companies access intimate behavioral data beyond social media through credit card transactions, GPS signals from phones, browsing histories, cookies, fitness trackers, smart home devices like Roombas with cameras, and facial recognition from public surveillance cameras—creating comprehensive profiles even without social media presence. - **Privacy-Preserving Technology:** Federated learning allows AI models to train on user devices locally, sending only updated intelligence back to companies rather than raw data, as Apple implements with Siri—enabling personalization without centralized data storage while reducing breach risks and compliance costs for businesses. → NOTABLE MOMENT Sandra demonstrates how a Roomba vacuum cleaner photographed a woman on the toilet and leaked the image online through contractors in Venezuela, revealing how seemingly innocuous smart devices collect intimate data users never intended to share when accepting vague terms and conditions. 💼 SPONSORS [{"name": "Donors Choose", "url": "https://donorschoose.org/local"}, {"name": "Capital One", "url": "https://capital1.com"}, {"name": "Fisher Investments", "url": "https://fisherinvestments.com"}, {"name": "Adobe", "url": "https://adobe.com/do-that-with-acrobat"}, {"name": "REI Co-op", "url": "https://rei.com"}, {"name": "Superhuman", "url": "https://superhuman.com/podcast"}, {"name": "Mint Mobile", "url": "https://mintmobile.com/switch"}, {"name": "Vital Farms", "url": "https://vitalfarms.com"}] 🏷️ Psychological Targeting, Data Privacy, Federated Learning, AI Profiling