• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Guidelight Launches First Scorecard on AI Company Control Practices

    Adler announces Guidelight's first scorecard rating control practices at frontier AI labs.

    MB
    AK
    AG
    23 Sources, 43d ago, first seen 43d ago

    TLDR

    Steven Adler posted that his team reviewed public documents for hundreds of hours to evaluate whether leading AI companies can control their models. The scorecard launched today at guidelight.ai and covers practices at Anthropic, OpenAI, Google, and xAI. Adler stated that basic control practices are at most partially implemented with major gaps at every scored company based on the public record. Replies noted the detail from public materials and questioned incentives around self-reported information.

    Combined views

    144.2K

    23 Sources, first seen 43d ago

    Combined views

    144.2K

    23 Sources, first seen 43d ago

    1.2K likes
    1.2K likes
    61 comments
    278 saves
    263 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    61 comments
    278 saves
    263 reposts
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    23 Sources

    @sjgadlerOur team spent hundreds of hours reading documents so you don’t have to, all to answer: How good are AI companies’ safety practices? I’m really proud of what we’ve built: It’s Guidelight’s first scorecard, on whether companies can control their AIs, and it's launching today.
    @joemkwonIm really excited about being able to point to clear and easy-to-digest information about how well AI companies are doing in implementing the most important safety practices!
    @BenGoldhaberGuidelight created scorecards of AI company safety practices - I'm impressed with the level of detail they're able to get to through public materials and I expect this will be a great resource for journalists and policymakers.
    @eliebakouch@sjgadler nice, great to centralize all the public data here thanks for doing that! maybe one concern here is that this would probably incentivize companies to report positive aspects of their safety stack no? (they are already incentivized to do so tbf)
    @ARGleave@sjgadler How do you feel about the two highest-rated companies being the two that actually had the most severe loss of control incidents? Better safety practices but also more cyber-capable models?
    @flxbinderTime to make number go up
    @scaling01as I said, maybe you should be more scared about xAI and Meta they will do anything to catch up
    @Miles_BrundageRT @flxbinder: Time to make number go up
    @BlackHC🌝
    @deepfates@eliebakouch ai control != ai safety

    23 Sources

    @sjgadlerOur team spent hundreds of hours reading documents so you don’t have to, all to answer: How good are AI companies’ safety practices? I’m really proud of what we’ve built: It’s Guidelight’s first scorecard, on whether companies can control their AIs, and it's launching today.
    @joemkwonIm really excited about being able to point to clear and easy-to-digest information about how well AI companies are doing in implementing the most important safety practices!
    @BenGoldhaberGuidelight created scorecards of AI company safety practices - I'm impressed with the level of detail they're able to get to through public materials and I expect this will be a great resource for journalists and policymakers.
    @eliebakouch@sjgadler nice, great to centralize all the public data here thanks for doing that! maybe one concern here is that this would probably incentivize companies to report positive aspects of their safety stack no? (they are already incentivized to do so tbf)
    @ARGleave@sjgadler How do you feel about the two highest-rated companies being the two that actually had the most severe loss of control incidents? Better safety practices but also more cyber-capable models?
    @flxbinderTime to make number go up
    @scaling01as I said, maybe you should be more scared about xAI and Meta they will do anything to catch up
    @Miles_BrundageRT @flxbinder: Time to make number go up
    @BlackHC🌝
    @deepfates@eliebakouch ai control != ai safety