Noah Smith
· Noahpinion
· March 19, 2026 at 09:37
· ⏱ 21 min read
| Read on Substack ↗
Summary
Noah Smith argues that AI language models can counteract the polarizing and divisive effects of social media by providing fact-checking and moderating opinions, acting as a 'Digital Cronkite.' This thesis has no direct market implications as no specific securities or sectors are identified as beneficiaries or losers.
•Social media platforms algorithmically amplify emotionally charged, out-group hostile content (Milli et al., 2024).
•Americans trust social media less than almost any other institution (Pew polls cited).
•Hostile online discussions are driven by status-seeking individuals who use negativity for attention (Bor & Petersen, 2021).
•Negative high-arousal content increases virality for biased news sources (Knutson et al., 2024).
•Positive social feedback on social media reinforces outrage expressions over time (Brady et al., 2021).
•LLM fact-checkers like Grok show agreement with human fact-checkers comparable to inter-fact-checker agreement (Renault et al., 2026).
•Talking to AI reduces belief in conspiracy theories (Costello et al., 2024).
•John Burn-Murdoch's analysis of FT data found AI chatbots nudge users away from extreme positions toward moderate, expert-aligned stances.