Daily Mail PH

Friday, November 3, 2023

[New post] AI Bot Transparency Index

Site logo image robertmcgrath posted: " Besides not working all that well, with a sky-high hype-to-performance ratio, this generation of large language models is also remarkable for low standards of documentation—even by the pitiful standards of the Internet. Even government regulators, who" Robert McGrath's Blog

AI Bot Transparency Index

robertmcgrath

Nov 3

Besides not working all that well, with a sky-high hype-to-performance ratio, this generation of large language models is also remarkable for low standards of documentation—even by the pitiful standards of the Internet.

Even government regulators, who are struggling to get a clue what to do, have figured this out.  We need to know what is in these beasts, where they come from, what they do, and what they are being used for.

Responding to this demand, researchers at Stanford assembled a report card for this years LLMs [1].

No surprise—everybody fails [2].

The report is based on open information, though the responsible parties had opportunity to add or correct the record.  And, for the record, Musk's "OpenAI", may or may not be "AI", but it sure isn't "open".

One of the researchers notes that in recent years, "[a]s the impact goes up, the transparency of these models and companies goes down," (Rishi Bommasani of CRFM quoted in [2]).

The report card covers the obvious things, like software specs, provenance of the training data, training methods, and so on.  It also includes "downstream" issues, such as access, access policies, and "impact". 

One of the interesting input variables is how human labor is used.  These large language models are "refined" by human supervisors.  For instance, there has been considerable hype about the supposed "guardrails" on ChatGPT.  These are basically human interventions to suppress some dangerously crazy results.

We know that these models are pretty much useless without human "tuning".  And we know that this tuning has a strong effect on the results, from the differences in versions of the same model.

So, it is extremely relevant to ask who are these humans, and what are they doing? 

The researchers note that it is widely believed that many of these humans are remote workers in low wage areas, such as Kenya.  But no one outside the companies really knows.


Is anything going to change?

I'm not holding my breath. 

With reports that OpenAI—which hasn't been "open" for years—is preparing a deal that will value the company at $80 billion, we can be sure they ain't gonna' be telling anybody anything anytime soon.

Sigh.


  1. Rishi Bommasani, Kevin Klyman, Shayne Longpre, Sayash Kapoor, Nestor Maslej, Betty Xiong, Daniel Zhang, and Percy Liang, The Foundation Model Transparency Index. Center for Research on Foundation Models (CRFM), Stanford, 2023. https://crfm.stanford.edu/fmti/
  2. Eliza Strickland, Top AI Shops Fail Transparency Test, in IEEE Spectrum - Artificial Intelligence, October 23, 2023. https://spectrum.ieee.org/ai-ethics
Comment
Like
Tip icon image You can also reply to this email to leave a comment.

Manage your email settings or unsubscribe.

Trouble clicking? Copy and paste this URL into your browser:
https://robertmcgrath.wordpress.com/2023/11/03/ai-bot-transparency-index/

WordPress.com and Jetpack Logos

Get the Jetpack app to use Reader anywhere, anytime

Follow your favorite sites, save posts to read later, and get real-time notifications for likes and comments.

Download Jetpack on Google Play Download Jetpack from the App Store
WordPress.com on Twitter WordPress.com on Facebook WordPress.com on Instagram WordPress.com on YouTube
WordPress.com Logo and Wordmark title=

Automattic, Inc. - 60 29th St. #343, San Francisco, CA 94110  

at November 03, 2023
Email ThisBlogThis!Share to XShare to FacebookShare to Pinterest

No comments:

Post a Comment

Newer Post Older Post Home
Subscribe to: Post Comments (Atom)

CG BOSS Quick check — should I keep you on this list?

You subscribed to CGBOSS in the past and Thank you  ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏...

  • [New post] Achieve Data Sovereignty through Omnisphere
    Crypto Breaking News posted: "Web 3.0 is one of the biggest buzzwords flying around the world of social media this year. An...
  • [New post] Tuesday’s politics thread is trying to stay positive.
    SheleetaHam posted: " Even though I just finished the latest Opening Arguments podcast about how Roe v. Wade is toast, and ...
  • How can the Rappler app be better? We'd like to know what you think!
    Hi daily! Have you downloaded the Rappler app? We'd love to hear about ...

Search This Blog

  • Home

About Me

Daily Newsletters PH
View my complete profile

Report Abuse

Labels

  • Last Minute Online News

Blog Archive

  • February 2026 (1)
  • January 2026 (7)
  • December 2025 (8)
  • November 2025 (4)
  • October 2025 (2)
  • September 2025 (1)
  • August 2025 (2)
  • July 2025 (5)
  • June 2025 (3)
  • May 2025 (2)
  • April 2025 (2)
  • February 2025 (2)
  • December 2024 (1)
  • October 2024 (2)
  • September 2024 (1459)
  • August 2024 (1360)
  • July 2024 (1614)
  • June 2024 (1394)
  • May 2024 (1376)
  • April 2024 (1440)
  • March 2024 (1688)
  • February 2024 (2833)
  • January 2024 (3130)
  • December 2023 (3057)
  • November 2023 (2826)
  • October 2023 (2228)
  • September 2023 (2118)
  • August 2023 (2611)
  • July 2023 (2736)
  • June 2023 (2844)
  • May 2023 (2749)
  • April 2023 (2407)
  • March 2023 (2810)
  • February 2023 (2508)
  • January 2023 (3052)
  • December 2022 (2844)
  • November 2022 (2673)
  • October 2022 (2196)
  • September 2022 (1973)
  • August 2022 (2306)
  • July 2022 (2294)
  • June 2022 (2363)
  • May 2022 (2299)
  • April 2022 (2233)
  • March 2022 (1993)
  • February 2022 (1358)
  • January 2022 (1323)
  • December 2021 (2064)
  • November 2021 (3141)
  • October 2021 (3240)
  • September 2021 (3135)
  • August 2021 (1782)
  • May 2021 (136)
  • April 2021 (294)
Simple theme. Powered by Blogger.