Is America Losing the Ability to Think Critically?
A parallel crisis is unfolding in classrooms and living rooms alike. NAEP and PISA data show reading proficiency in steady decline—only about 35% of high school seniors can now perform basic textual analysis—while The Atlantic and WSJ both point to hasty reading habits, AI reliance, and classroom distraction as culprits [4][5][6]. The result, critics argue, is a generation more prone to dismissing arguments with a shrug ("you used AI") than actually engaging with them.
Education reformers argue the failure is systemic: schools have stopped teaching comprehension, logical fallacies, and source evaluation, leaving adults defenseless against memes and outrage bait. Others push back, framing this as a cultural rather than institutional failure—pointing to book bans, identity-driven curriculum fights, and a broader devaluation of deep reading, with AI serving as a crutch that reduces cognitive struggle rather than causing it outright.
Either way, the consequence is the same: a public less equipped to evaluate competing claims, which feeds directly into the polarization crisis above.
Who Should Police Speech—Platforms, Courts, or AI?
As content moderation battles continue, courts have affirmed platforms' right to curate algorithmically as protected speech under Moody v. NetChoice, but the debate is shifting toward a thornier question: should AI systems themselves act as arbiters of acceptable speech? [7][8][9]
Free speech advocates warn that AI moderators—especially if nudged by government pressure—risk viewpoint discrimination without the legal safeguards that exist for something like copyright's fair-use doctrine. Safety-focused critics counter that unmoderated spaces enable real harms—hate speech, disinformation, incitement—and note that laws like New York's hate speech statute have been partially upheld, suggesting some regulation is constitutionally viable. A middle position holds that AI could serve as a neutral arbiter, but only if moderation criteria avoid political viewpoint-based judgments—a big "if" in a polarized environment.
AI's Existential Risks Collide With Data Center Backlash
Silicon Valley and Washington are increasingly at odds over how fast to build. Former Anthropic researchers have warned about self-improving AI systems outpacing human oversight, while Trump has framed rapid buildout as essential to beating China [10]. Meanwhile, everyday Americans are turning against the infrastructure this all requires: opposition to nearby data centers has jumped from 28% to 41%, driven by concerns over electricity and water use [11].
This has produced an unusual bipartisan moment. Progressives like Sanders are pushing moratoriums, taxes, and bans on "superintelligence" research, while some Republicans are joining in—not over existential risk, but over local energy costs [12]. Accelerationists counter that safety panics echo past technological scares, and that ceding ground now hands China a permanent advantage. The result is a rare issue where left-right coalitions form around shared anxiety, even if the underlying reasons diverge sharply.
The Bigger Picture
Today's stories share a common thread: the erosion of shared ground—whether it's a shared set of facts, a shared capacity to evaluate claims, shared rules for public speech, or a shared sense of how fast humanity should move on transformative technology. In each case, reasonable people land in genuinely different places not because one side is uninformed and the other enlightened, but because they're weighing real tradeoffs differently—innovation versus caution, safety versus liberty, national competitiveness versus local harm.
What's notable is how often these debates get flattened into caricature. The critical thinking crisis isn't simply "kids these days," and the AI safety debate isn't simply accelerationists versus doomers—both contain legitimate empirical claims and legitimate uncertainty. The healthiest response isn't picking a side reflexively, but doing the harder work of understanding why an intelligent, informed person might hold the opposing view.
Key takeaway: The through-line connecting media fragmentation, declining literacy, speech regulation, and AI policy is the same: productive disagreement requires shared facts, the skills to evaluate them, and enough humility to ask what the other side actually believes—before deciding they're wrong.
Sources
- https://www.thebulwark.com/p/how-maga-bewildered-voters-into-its-alternate-reality
- https://www.science.org/doi/10.1126/science.aei2409
- https://www.nytimes.com/2026/08/10/style/america-is-atomizing.html
- https://www.theatlantic.com/magazine/2026/08/reading-crisis-postliterate-age/687618/
- https://www.wsj.com/opinion/yes-children-really-are-getting-dumber-fa4cb20c
- https://bcgavel.com/2026/04/26/the-literacy-crisis-in-america/
- https://reason.com/volokh/2026/05/06/content-moderation-and-the-first-amendment/printer/
- https://cdt.org/event/ai-speech-and-the-first-amendment-where-are-the-constitutional-lines/
- https://cdt.org/insights/not-all-guardrails-are-created-equal-comparing-content-safety-and-copyright-filtering/
- https://www.nytimes.com/interactive/2026/09/15/technology/ai-regulation-risks-politicians-executives.html
- https://www.politico.com/news/2026/07/21/poll-data-centers-democrats-moratorium-01001799
- https://www.wired.com/story/candidates-are-signing-a-pact-promising-action-on-data-centers-and-ai-safety/