Daily Tech News Show We're More Sorted Than Polarized - DTNS 4916
5 snips
Dec 16, 2024 Andrea Jones-Rooy, a data science professor and host of the Behind the Data podcast, and Dr. Maryanne Garry, a cognitive psychology professor at the University of Waikato, tackle the pressing issue of political polarization. They explore the data behind the phenomenon, revealing how voting behaviors reflect ideological divides. Dr. Garry also discusses the persuasive power of large language models, highlighting the psychological factors that lead users to trust AI outputs, along with the potential risks of this misplaced confidence.
AI Snips
Chapters
Transcript
Episode notes
Sorted, Not Polarized
- Political polarization feels widespread, but data reveals a different story.
- We're more sorted than polarized, meaning our views align more consistently within party lines, eliminating common ground.
Persuasive LLMs
- LLMs can be persuasive due to their conversational nature, confidence, and how they mimic human thought processes.
- These qualities, while designed for approachability, can inadvertently mislead us.
Using LLMs Carefully
- Be aware of LLM limitations, even if tech-savvy.
- Triangulate information and question sources, including the chatbot itself.