DOOMBENCH.COM
Safety and alignment

OpenAI enables opt-in text watermarking with restricted detector access

OpenAI made textGrain watermarking available as a global API opt-in and opened applications for researcher detector access. It disclosed that editing and short passages can undermine detection. EU ChatGPT and Codex rollout remains planned, not completed.

0 comments · 0 votesOpen discussion

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

No comments yet. Start the discussion.

Send detailed or private feedback →
CURRENT ASSESSMENT · REVISION 1
AWAY FROM DOOM20confidence 60/100

Why it moved the index

Text provenance can support oversight of AI-generated manipulation; the modest magnitude and capped confidence reflect restricted detection access, evasion limits and an inferred rather than measured reduction in harm.

Assessment history

  1. R1
    Away 20 · confidence 60

    New text-watermarking deployment, distinct from earlier image/audio provenance and Anthropic's separate text controls.

    06 Oct 2026

Share this page

DoomBench social sharing card for OpenAI enables opt-in text watermarking with restricted detector access.

  1. DoomBench assesses “OpenAI enables opt-in text watermarking with restricted detector access” as evidence moving away from doom, with magnitude 20 and confidence 60 out of 100 in the safety and alignment category.

  2. The DoomBench assessment of “OpenAI enables opt-in text watermarking with restricted detector access” is based on reporting from OpenAI and records the editorial rationale, source quality, attribution, and revision history.

  3. DoomBench summarizes “OpenAI enables opt-in text watermarking with restricted detector access” as follows: OpenAI made textGrain watermarking available as a global API opt-in and opened applications for researcher detector access. It...

    https://www.doombench.com/news/openai-enables-opt-in-text-watermarking-with-restricted-detector-access-2026-10-05