Human resilience

OpenAI withdraws AI-text classifier for low accuracy

OpenAI removed its AI-written-text classifier after reporting low accuracy, including only 26% true-positive identification and a 9% false-positive rate on its challenge set.

0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

CURRENT ASSESSMENT · REVISION 1
TOWARD DOOM35confidence 96/100

Why it moved the index

Withdrawal documented the practical weakness of a proposed institutional detection control, leaving educators and publishers without a reliable text-origin safeguard.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 35 · confidence 96

    New July 2023 counterevidence on generative-text detection controls.

    12 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for OpenAI withdraws AI-text classifier for low accuracy.
  1. DoomBench assesses “OpenAI withdraws AI-text classifier for low accuracy” as evidence moving toward doom, with magnitude 35 and confidence 96 out of 100 in the human resilience category.

  2. The DoomBench assessment of “OpenAI withdraws AI-text classifier for low accuracy” is based on reporting from OpenAI and records the editorial rationale, source quality, attribution, and revision history.

  3. DoomBench summarizes “OpenAI withdraws AI-text classifier for low accuracy” as follows: OpenAI removed its AI-written-text classifier after reporting low accuracy, including only 26% true-positive identification and a 9% false-positive...

    https://www.doombench.com/news/openai-withdraws-ai-text-classifier-for-low-accuracy-2023-07-20