#453 — AI and the New Face of Antisemitism - Making Sense with Sam Harris Recap

Podcast: Making Sense with Sam Harris

Published: 2026-01-16

Duration: 22 min

Summary

In this episode, Sam Harris engages with Judea Pearl about the rising concerns of antisemitism in contemporary culture and the implications of artificial intelligence (AI) on this issue. They explore the limitations of current AI technologies and the potential dangers of achieving artificial general intelligence (AGI).

What Happened

Sam Harris welcomes Judea Pearl, who has returned to discuss his new book, 'Coexistence and Other Fighting Words.' Their conversation begins with Pearl's background, touching on his family's immigration to Israel in 1924, motivated by a traumatic incident of antisemitism. This personal history lays the groundwork for a broader discussion on cultural issues, particularly the alarming rise of antisemitism in both left and right political spectrums, and how Israel's status has been impacted on the global stage.

As they delve into the topic of artificial intelligence, Pearl shares his insights on the current advancements in large language models (LLMs). While acknowledging their impressive capabilities, he expresses skepticism about their potential to lead to AGI, citing mathematical limitations that cannot be overcome merely by scaling up data and computational power. Pearl emphasizes the importance of causal reasoning, pointing out that LLMs primarily summarize existing data without truly understanding the underlying causal relationships. This distinction is crucial when considering the implications of AI on society and cultural dynamics, especially in the context of rising antisemitism and misinformation.

Key Insights

Key Questions Answered

What motivated Judea Pearl's family to move to Israel?

Judea Pearl recounts a family legend about his grandfather, who was beaten by a Polish peasant and called a 'dirty Jew.' This traumatic experience prompted his grandfather to declare, 'Start packing, We are going to where we belong.' This incident underscores the historical context of antisemitism that influenced their migration in 1924.

What are Judea Pearl's views on the current state of AI?

Pearl expresses a nuanced perspective on AI, particularly current large language models. He acknowledges their impressive performance but believes that they do not lead us toward AGI. He asserts that they are limited by their reliance on summarizing data produced by human authors rather than engaging in true causal reasoning, which he argues is essential for understanding complex systems and making informed decisions.

How does Judea Pearl differentiate between correlation and causation?

Pearl emphasizes that one cannot derive causation from correlation, a fundamental principle in statistics. He explains that while LLMs can analyze data, they lack the ability to perform interventions that would allow them to understand causal relationships. This limitation is critical when considering the implications of AI in decision-making processes and understanding societal issues like antisemitism.

What potential dangers does Judea Pearl foresee with advanced AI?

Pearl expresses concern about the potential for a truly AGI system to pose a danger to humanity. He acknowledges that, while current AI technologies do not threaten free will or consciousness, the future possibilities of AI gaining such capabilities raise alarms. He notes that the dangers of misinformation and misalignment in AI systems are already apparent, prompting concerns about their broader societal impacts.

What role does Judea Pearl believe AI plays in the rise of antisemitism?

While Pearl and Harris discuss the broader implications of AI, Pearl highlights that the rise of antisemitism is a cultural issue exacerbated by misinformation and narratives propagated through various media, including AI-generated content. He stresses the importance of understanding these dynamics to combat hate and misinformation effectively, which requires a robust understanding of causality that current AI models are not equipped to provide.