The AI Trust Gap: Users Rely on Chatbots While Doubting Their Accuracy

The survey reveals that convenience often overrides caution, with 47.3% of respondents stating they stop questioning an answer simply because it sounds plausible. This passive acceptance persists despite one in four users admitting they have already applied AI-generated advice to high-stakes life decisions involving personal health, finances, or relationships.

F-Secure identifies a concerning trend where frequent users lower their guard, with 59.7% of power users accepting answers without verification. Many attempt to mitigate risk by using one AI to fact-check another, a practice experts warn creates a false sense of security by asking the technology to audit its own output. Amel Bourdocen, a senior researcher at F-Secure, argues that shifting the burden of verification onto users is a flawed strategy. Instead, she suggests that developers must design systems capable of signaling their own uncertainty, rather than presenting potentially incorrect information with artificial confidence.