Nieman Lab: MIT study finds AI reliance erodes news credibility judgment

Researchers at the MIT Media Lab published a study in June 2026 finding that participants who used AI chatbots to assess news credibility saw their independent detection skills deteriorate significantly once AI assistance was removed. The paper was presented at the 2026 CHI Conference on Human Factors in Computing Systems.

The study tracked 67 participants over four weeks as they assessed the credibility of news headline-image pairs. During phases when AI assistance was available, participants’ accuracy in identifying misinformation improved by 21%. But when AI was removed in the final phase, performance dropped 15.3 percentage points below where participants had started — worse than if they had never used AI at all. Roughly a quarter of participants reported feeling that they were improving even as their measured performance declined.

The researchers frame this as an “AI dependency paradox”: the AI helped in the moment, but the pattern of offloading judgment to it prevented participants from developing the independent skills needed to maintain that accuracy without it. The study connects this to broader research on cognitive offloading — the same mechanism observed with GPS navigation and calculators.

The distinction the paper draws between AI as coach and AI as crutch is the most actionable part for newsrooms. An AI system that prompts users to reason through their own assessment before offering a verdict supports skill development. A system that simply delivers an answer replaces the reasoning step. For editors and fact-checkers evaluating AI verification tools, this study offers a concrete design criterion: does the tool require users to engage their own judgment, or does it substitute for it?