MIT's Eric So warns AI overreliance erodes cognitive skills in adults and youth
Source: Fox Business · All Fox Business reports
Unlock the full scoreboard
Letter grade, factuality, lean, and rationales — free with registration. No card required.
See grades free How grading works
Already have an account? Sign in. The full report below is free to read.
Disagree with this grade or political lean?
Flagging is open to every reader with a free account. Sign in or create one to dispute this report.
Topics in this report
Summary
The Fox Business segment interviews MIT Sloan professor Eric So about his new book "The Collision: What AI Does to Us," in which he argues that overreliance on AI acts as a crutch, bypassing the mental struggle needed to build foundational skills. So warns of adults losing existing capabilities and young people never developing them, creating a reinforcing cycle of dependency. He shares personal habits that prompted the book and recommends open discussions on drawbacks, focusing on process over outcomes, and taking periodic breaks from AI. The interview then shifts to the White House Super Intelligence Task Force's first meeting the previous day, asking whether legal liability suffices as regulation for catastrophic risks. So says liability is important but insufficient alone, stressing the need to preserve human judgment for monitoring and intervention.
The segment relies entirely on So as the named expert, referencing his research and book without anonymous sources or opposing views. Graphics are not mentioned; the throughline connects personal cognitive risks to broader societal and regulatory needs for human oversight.
Editorial Assessment
The broadcast accurately conveys So's core thesis from his recently published book, which synthesizes cognitive science, behavioral research, and his observations, including an Anthropic study on student AI queries. Claims hold up against supporting studies showing performance drops when AI is removed after use, though the field includes mixed results where guided AI can aid learning. Viewers may miss broader context: many studies show productivity gains when AI augments rather than replaces thinking, and long-term evidence on "cognitive atrophy" remains emerging rather than conclusive. The framing leans toward caution and personal responsibility, potentially skewing perception that AI's primary effect is weakening humans without equal time on upsides like accelerated innovation. The brief pivot to the task force accurately notes its recent inaugural meeting but treats liability in isolation from the administration's broader industry-accord and risk-report approach. Overall, a thoughtful expert interview that surfaces real concerns but could benefit from counterbalancing data.
Key Moments
Overusing AI without engaging in thought and struggle undermines foundational capability, mastery, and creates reinforcing cycle where people lose comfort doing tasks without it.
Core thesis of So's book 'The Collision,' echoed in MIT News Q&A and Sloan coverage; supported by studies on cognitive offloading and persistence loss after brief AI use.
Adults risk losing skills they once had; young people risk never developing them at all.
Directly from So in the interview and book; aligns with MIT reports on student 'cognitive surrender' and research on deskilling in education.
Practical steps include open discussions on drawbacks, focus on process not just outcome, and taking breaks from AI to assess lost capabilities.
So's personal practice and recommendations detailed in his MIT interview and book companion workbook.
White House Super Intelligence Task Force held its first meeting yesterday; legal liability is an important but incomplete part of AI regulation, especially for catastrophic risks where preserved human judgment is needed to monitor and intervene.
Task force met Oct. 15, 2026, per multiple reports; So's view consistent with expert proposals on liability for frontier AI risks, though administration emphasizes self-regulation via accords.
Notable Concerns
- Single-expert sourcing without counterpoints from AI optimists or conflicting studies
- Limited context on how AI use context (e.g., as tutor vs. answer provider) affects outcomes
Sources Consulted
- ‘The Collision: What AI Does to Us’ | MIT Sloan
- Q&A: The “collision” between AI, your brain, and society | MIT News
- White House ‘Super Intelligence’ Task Force to hold its first meeting - POLITICO
- White House’s AI task force gets to work | Semafor
- THE COLLISION | Kirkus Reviews
- Is AI making us stupid through cognitive offloading? New review explores the evidence
- Study explores how AI use affects performance on math and reasoning tasks - Berkeley News
- Overreliance on AI: Literature Review - Microsoft Research