AI Threat Reveals an Uncomfortable Truth: Human Research Findings Are Often False

RexDouglass · x · 2026-08-05

@RexDouglass argues that the perceived "threat" of AI (like generating massive amounts of disinformation) has merely made people realize what decades of metascience already proven: human-generated academic slop is equally prevalent and unreliable.

The quoted tweet echoes this, noting this isn't unique to AI. A classic 2005 paper already stated that most research findings are false. The author criticizes the academic culture of pushing papers to "tell a story," arguing that "story sells better than truth," which leads researchers to discard data that doesn't fit the narrative, resulting in biased outcomes.

Original post →

More from AGI Musings

AGI Musings channel →