When AI fabricates your quotes

A recent New York Times investigation revealed that a high-profile book, The Future of Truth, included fabricated quotes attributed to data journalist Meredith Broussard—despite her never having written them. Broussard, a leading critic of algorithmic bias, traced the misattribution to a 2023 interview where she made a similar point in conversation, which an AI system then rephrased and falsely presented as a book quote. This incident exposes a growing crisis in epistemic trust: AI doesn’t just generate content, it fabricates credibility. Broussard argues that the real danger isn’t AI lying outright, but its ability to sound plausible—so convincing that users mistake it for truth, especially when they’re not experts in the subject. She warns that over-reliance on AI erodes critical thinking, particularly among younger users who lack the life experience to spot generic advice when it’s disguised as insight. Even more alarming, she notes that students now come to her class already aware of AI’s flaws—yet many still refuse to use it, recognizing that relying on it undermines learning. Meanwhile, Google’s decline and the rise of privacy-focused search engines like DuckDuckGo suggest a growing public fatigue with AI-driven summaries. The episode underscores a pivotal shift: we’re no longer just using AI as a tool, but risking it as a substitute for judgment, memory, and expertise.