

Algorithm Shrugged
36 snips May 29, 2025
Zeve Sanderson, the Founding Executive Director of the NYU Center for Social Media and Politics, joins the discussion to tackle the accountability of tech companies in cases of radicalization linked to online platforms. They delve into the persuasive power of large language models and examine various recent laws aimed at regulating AI, such as California's AI Transparency Act. Carter Huffman, CTO of Modulate, shares insights on innovative voice technology designed for fraud detection, highlighting its role in countering sophisticated impersonation tactics.
AI Snips
Chapters
Transcript
Episode notes
Buffalo Shooter's Online Radicalization
- The shooter in the 2022 Buffalo attack was radicalized by online racist content and live streamed the event.
- This case raised questions on tech platforms’ liability for algorithmic radicalization beyond just algorithm influence.
Algorithmic Radicalization Overstated
- Research shows extremist content consumption on YouTube is driven mainly by user subscriptions and external links, not algorithmic recommendations.
- This challenges the common narrative that algorithms primarily radicalize users on platforms like YouTube.
Demand Drives Extremist Content
- There is a strong demand-side factor where users with prior biases seek extremist content rather than being passively radicalized by algorithms.
- Focusing only on algorithms misses deeper questions about user motivations and platform responsibilities.