conv.

All stories
AIQuiet 5d · day 8

Academics Set Minimum Standards for Embedded AI Evaluators

An open letter from the AI Evaluator Forum, signed by academics including a 2024 Nobel physics laureate, lays out baseline conditions for third-party evaluators embedded at frontier AI companies.

newest coverage 5d agosummary written 7d ago5 pieces5 likes & upvotes1 comments Hacker NewsRedditNewswiresMastodon

What to know

  • The letter sets minimum conditions—scientific objectivity, transparency, independence, and protection from interference—for third-party evaluators embedded at frontier AI companies.
  • Signatories include a 2024 Nobel physics laureate and researchers from MIT, Cambridge, Carnegie Mellon, Berkeley, Wisconsin-Madison, Uppsala and UNC Chapel Hill, all signing in a personal capacity.
  • The letter responds to recent, unspecified announcements from frontier AI companies about embedding evaluators, and cites the existing AEF-1 standard as an early model.
  • It frames embedded evaluation as a complement to, not a substitute for, broader public transparency and independent researcher access.

“To be credible, embedded third-party evaluations must have scientific objectivity, transparency, independence, and robust protections against interference from the evaluated companies”

AI Evaluator Forum open letter, Signatory academics · aievaluatorforum.org ↗ · Sep 17

AI Evaluator Forum Organizer of the open letterNobel Laureate in Physics (2024) Signatory, University Professor Emeritus, University of TorontoFrontier AI companies (unnamed, footnoted 1–5) Subjects of the letter's demands

How it unfolded 3 developments, newest first · click a bar or a number to jump articlesposts

Peak 1 piece in two hours at Sep 18, 11 AM; 5 pieces over 8 days (1 article · 4 posts) Sep 18, 11 AM — 1 piece · 1 post — Hacker News 1Sep 18, 1 PM — quietSep 18, 3 PM — 1 piece · 1 article — Newswires 1Sep 18, 5 PM — 1 piece · 1 post — Reddit 1Sep 18, 7 PM — quietSep 18, 9 PM — quietSep 18, 11 PM — quietSep 19, 1 AM — quietSep 19, 3 AM — quietSep 19, 5 AM — quietSep 19, 7 AM — quietSep 19, 9 AM — quietSep 19, 11 AM — quietSep 19, 1 PM — quietSep 19, 3 PM — quietSep 19, 5 PM — quietSep 19, 7 PM — quietSep 19, 9 PM — quietSep 19, 11 PM — quietSep 20, 1 AM — quietSep 20, 3 AM — quietSep 20, 5 AM — quietSep 20, 7 AM — quietSep 20, 9 AM — quietSep 20, 11 AM — quietSep 20, 1 PM — quietSep 20, 3 PM — quietSep 20, 5 PM — quietSep 20, 7 PM — quietSep 20, 9 PM — quietSep 20, 11 PM — quietSep 21, 1 AM — quietSep 21, 3 AM — 1 piece · 1 post — Mastodon 1Sep 21, 5 AM — quietSep 21, 7 AM — 1 piece · 1 post — Mastodon 1Sep 21, 9 AM — quietSep 21, 11 AM — quietSep 21, 1 PM — quietSep 21, 3 PM — quietSep 21, 5 PM — quietSep 21, 7 PM — quietSep 21, 9 PM — quietSep 21, 11 PM — quietSep 22, 1 AM — quietSep 22, 3 AM — quietSep 22, 5 AM — quietSep 22, 7 AM — quietSep 22, 9 AM — quietSep 22, 11 AM — quietSep 22, 1 PM — quietSep 22, 3 PM — quietSep 22, 5 PM — quietSep 22, 7 PM — quietSep 22, 9 PM — quietSep 22, 11 PM — quietSep 23, 1 AM — quietSep 23, 3 AM — quietSep 23, 5 AM — quietSep 23, 7 AM — quietSep 23, 9 AM — quietSep 23, 11 AM — quietSep 23, 1 PM — quietSep 23, 3 PM — quietSep 23, 5 PM — quietSep 23, 7 PM — quietSep 23, 9 PM — quietSep 23, 11 PM — quietSep 24, 1 AM — quietSep 24, 3 AM — quietSep 24, 5 AM — quietSep 24, 7 AM — quietSep 24, 9 AM — quietSep 24, 11 AM — quietSep 24, 1 PM — quietSep 24, 3 PM — quietSep 24, 5 PM — quietSep 24, 7 PM — quietSep 24, 9 PM — quietSep 24, 11 PM — quietYesterday, 1 AM — quietYesterday, 3 AM — quietYesterday, 5 AM — quietYesterday, 7 AM — quietYesterday, 9 AM — quietYesterday, 11 AM — quietYesterday, 1 PM — quietYesterday, 3 PM — quietYesterday, 5 PM — quietYesterday, 7 PM — quietYesterday, 9 PM — quietYesterday, 11 PM — quietToday, 1 AM — quietToday, 3 AM — quietToday, 5 AM — quietToday, 7 AM — quietToday, 9 AM — quietToday, 11 AM — quietToday, 1 PM — quiet 1–3
Sep 19Sep 20Sep 21Sep 22Sep 23Sep 24yesterdaynow · 3:37 PM ET
  1. 3

    Letter circulates on Reddit and Hacker News

    The open letter is cross-posted to r/technology-style aggregators and Hacker News, drawing minimal engagement (a handful of votes, one comment) as of the latest data.

    “Embedded evaluations cannot address all oversight needs and should be treated as a complement to, rather than a replacement for, broader efforts by frontier AI companies to expand external oversight…”
    — AI Evaluator Forum open letter, Signatory academics · source
    • clarinette@mastodon.online

      # AI # LLM independent evaluators needed https:// aievaluatorforum.org/initiativ es/embedded-evaluation-letter

      clarinette@mastodon.onlineMastodon5d agoview on Mastodon ↗
    2 posts in this stretch →
  2. 2

    AI Evaluator Forum reposts letter with full signatory list

    The organizing group's own site republishes the letter, listing signatories by affiliation only (not by name), and references the AEF-1 standard as an example of codified conditions already seeing early adoption.

    1. 1 outlet Minimum Conditions for Embedding Evaluators

      first by AI Evaluator Forum, 7d ago

  3. 1

    Academics publish letter on minimum conditions for embedded AI evaluators

    The AI Evaluator Forum releases an open letter signed by researchers from MIT, Cambridge, Carnegie Mellon, Berkeley, Wisconsin-Madison, Uppsala, UNC Chapel Hill and a 2024 Nobel physics laureate, setting out minimum conditions—scientific objectivity, transparency, independence, and protection from interference—that third-party evaluators embedded within frontier AI companies must meet to be credible.

    “We believe that all frontier AI companies should embed evaluators to independently assess AI risks, including evaluating the systems themselves and any significant incidents of real-world harm…”
    — AI Evaluator Forum open letter

What people are saying 0 voices from 0 sites · best of 2 · verbatim