Jay Van Bavel, a psychology professor at NYU, makes a case that should be obvious but somehow still isn't mainstream: telling people the world is ending makes them scroll, not act. The argument is backed by hard experimental data, not vibes. The centrepiece is a study of 59,440 participants across 63 countries, testing 11 climate messages crafted by experts. The doom-laden passage — describing terrors "within the lifetime of a young adult today" — was the single most effective message for social media sharing. It was simultaneously the worst performer for inspiring real action: tree planting, political engagement, lifestyle change. The mechanism is textbook fear-appeal psychology: one-shot fear can mobilize, but sustained dread without efficacy cues produces learned helplessness. Van Bavel and colleagues also analysed 22,743 A/B headline tests from Upworthy.com, finding each additional negative word in a headline increased clicks by 2.3%. This is the engine under the hood: platforms optimize for engagement, engagement selects for negativity, and negativity selects against the behavioral responses that could actually address the crises being reported. The information ecosystem doesn't just describe doom — it manufactures the paralysis that makes doom more likely. The productive counter-evidence is specific. The scientific consensus message — "99% of expert climate scientists agree" — outperformed doom on every outcome except internet sharing. PhD student Danielle Goldwert's research found that framing climate action in terms of emotional benefits (hope, connection) significantly increased political participation, advocacy, donation, and lifestyle change. The signal is consistent: efficacy messaging beats fear messaging on every metric that matters except virality. The Doomsday Clock, created in 1947 by the Bulletin of the Atomic Scientists, gets a cameo as the original doom-communication device — now set at 85 seconds to midnight, down from the original seven minutes. Van Bavel positions it as an early example of a persistent tension: crisis communication that succeeds at grabbing attention but may fail at its stated purpose of inspiring action. The uncomfortable implication cuts across every major threat domain — climate, AI safety, geopolitics. Former Anthropic researcher Jacob Coxon's recent warnings about existential AI risk face the same trap: if the message produces paralysis rather than engagement with governance and technical safety work, the warning may be counterproductive regardless of its accuracy. The problem isn't that the risks are exaggerated. The problem is that the messaging architecture selects for the emotional response least likely to produce mitigation. What Van Bavel is really describing is a market failure in the attention economy. Doom is cheap to produce, easy to share, and monetizable through engagement metrics. Hope-plus-efficacy messaging is harder to craft, less viral, and requires specificity about what actions are available. The asymmetry is structural, not accidental, and it means the information ecosystem is systematically subsidizing paralysis over agency.
