DOOMBENCH.COM
Autonomy and agency

OpenAI WebGPT browses the web and outperforms human demonstrators

WebGPT autonomously issued searches, followed links, scrolled pages, gathered citations, and produced long-form answers. Its best 175B system was preferred over human demonstrations 56 percent of the time, and later motivated deployed ChatGPT browsing plugins.

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
TOWARD DOOM41confidence 93/100

Why it moved the index

This was a general language model gaining consequential external-tool autonomy that directly influenced a later released browsing system.

Assessment history

  1. R1
    Toward 41 · confidence 93

    New original research result with exact official RSS clock, exact 175B system identity, and separate primary practical-impact evidence from ChatGPT plugins.

    12 Aug 2026

Share this page

DoomBench social sharing card for OpenAI WebGPT browses the web and outperforms human demonstrators.

  1. DoomBench assesses “OpenAI WebGPT browses the web and outperforms human demonstrators” as evidence moving toward doom, with magnitude 41 and confidence 93 out of 100 in the autonomy and agency category.

  2. The DoomBench assessment of “OpenAI WebGPT browses the web and outperforms human demonstrators” is based on reporting from OpenAI and records the editorial rationale, source quality, attribution, and revision history.

  3. DoomBench summarizes “OpenAI WebGPT browses the web and outperforms human demonstrators” as follows: WebGPT autonomously issued searches, followed links, scrolled pages, gathered citations, and produced long-form answers. Its best 175B...

    https://www.doombench.com/news/openai-webgpt-browses-the-web-and-outperforms-human-demonstrators-2021-12-16