NOISEReport
Coverage history

Assembled by NOISE.

Report / Tech

Published by NOISE

When AI search helps, and when it gets it wrong

From Morning NOISE.

As chatbots like ChatGPT, Claude, and Gemini increasingly search the web for users, a Wired guide by David Nield outlines when the approach pays off and when it does not. AI search works best for widely documented queries and complex requests that require synthesizing multiple sources, but it can confidently present wrong answers when accurate sources are scarce or the query is misinterpreted, per Stanford research cited in the piece. Traditional search remains safer for precise details, breaking news, and high-stakes topics like health, law, and finances, and UC San Diego researchers have found AI tools may favor topically relevant sources over authoritative ones. The takeaway: users should double-check the sources behind any AI-generated answer.

Saved versions

These dates record changes to NOISE’s copy. They are not event dates.

  1. Saved copyCurrentIn saved edition