PulseAugur
LIVE 12:25:39
tool · [1 source] ·
0
tool

ChatGPT struggles with consistent carb estimates, showing AI unreliability

A user found that ChatGPT provided significantly different carb estimates for the same lunch photo when asked repeatedly. Across 500 queries, the AI's responses varied widely, indicating a lack of consistent output for a critical task. This inconsistency raises concerns about the reliability of AI in applications where precise and stable answers are necessary, such as dietary management. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Highlights potential unreliability of LLMs for precise, repeated tasks, impacting user trust in AI-driven recommendations.

RANK_REASON User-generated report of inconsistent AI output in a specific application.

Read on Mastodon — sigmoid.social →

COVERAGE [1]

  1. Mastodon — sigmoid.social TIER_1 · [email protected] ·

    "Ask ChatGPT to estimate the carbs in your lunch. Now ask it again. And again. Five hundred times." "You'd expect the same answer each time. It's the same photo

    "Ask ChatGPT to estimate the carbs in your lunch. Now ask it again. And again. Five hundred times." "You'd expect the same answer each time. It's the same photo, the same model, the same question. But you won't get the same answer. Not even close -- and the differences are large …