I asked 6 popular AIs these trick questions, and it was a game of hallucination roulette
SMRTR summary
A technology writer tested six popular AI tools with trick questions designed to expose hallucinations and factual errors, finding that while accuracy has generally improved, significant mistakes still occur inconsistently across platforms. Google Gemini fabricated book titles, Meta AI miscounted letters in "strawberry," and ChatGPT still hallucinates details about a famous fake legal case.
SMRTR provides this summary for quick context. The original article belongs to ZDNet.
Read the original article