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

OpenAI proposes shared global AI safety standards

OpenAI calls for coordinated evaluations, reporting and governance to improve safety. A post summarizing the proposal says it asks the U.S. to lead the effort.

Sam AltmanSA
Stella BidermanSB
Andrew CurranAC
10 Sources, 20d ago, first seen 20d ago

TLDR

The proposed U.S.-led framework would cover capability evaluations, risk assessment, human oversight and common incident reporting, according to a post describing OpenAI's proposal. The account says OpenAI wants standards to track how much research inside a lab is performed autonomously, with thresholds for immediate human review. It also says OpenAI warns that more autonomous research could become too opaque for humans to supervise effectively.

Combined views

644.9K

10 Sources, first seen 20d ago

6.5K likes1.1K comments1.6K saves525 reposts

Combined views

644.9K

10 Sources, first seen 20d ago

6.5K likes1.1K comments1.6K saves525 reposts

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Featured Source

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

10 Sources

Andrew Curran@AndrewCurran_OpenAI says that the United States should lead an effort to develop international standards for RSI, link the AI safety institutes into one network, and develop a system to track 'RSI-relevant' progress and the amount of autonomous research happening inside AI companies.20d
Rohan Paul@rohanpaul_aiOpenAI just released a proposal for global coordination on AI safety. It asks the U.S. to lead an international framework for capability evaluations, risk assessment, human oversight and common incident reporting. OpenAI says standards should track RSI-relevant capability gains and how much research inside a lab is performed autonomously, including thresholds for immediate human review. It warns that more autonomous research could become too opaque for humans to supervise effectively. The timing overlaps with separate U.S.-China talks on a proposed notification mechanism for AI incidents reaching national-security significance.20d
Sam Altman@samaPeople outside the AI labs should have a real say in how this technology develops, and a clear way to judge if it's happening safely. Standards should help prevent the concentration of power, including by making sure new companies and open-model companies can compete. They should also help countries and companies compare evidence and learn from failures. We think the US should lead this effort. Here is our proposal: https://openai.com/index/building-standards-next-phase-ai/19d
Brian Roemmele@BrianRoemmele@sama I analyzed and responded to Sam’s posting. Unfortunately what he is posting is not what it seems. Give it a read.19d
Stella Biderman@BlancheMinervaNotably this proposal contains none of: - People outside the labs having a real say it develops - a clear way to judge if it's happening safely. - Standards toprevent concentration of power - Aids to the competitive ability of open-model companies Maybe he posted the wrong one?18d
Daniel Kokotajlo@DKokotajloRT @trevposts: Per their post today about standards, OpenAI's plan for building AGI is to "build an automated AI researcher" and "iterate w…18d
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    10 Sources

    Andrew Curran@AndrewCurran_OpenAI says that the United States should lead an effort to develop international standards for RSI, link the AI safety institutes into one network, and develop a system to track 'RSI-relevant' progress and the amount of autonomous research happening inside AI companies.20d
    Rohan Paul@rohanpaul_aiOpenAI just released a proposal for global coordination on AI safety. It asks the U.S. to lead an international framework for capability evaluations, risk assessment, human oversight and common incident reporting. OpenAI says standards should track RSI-relevant capability gains and how much research inside a lab is performed autonomously, including thresholds for immediate human review. It warns that more autonomous research could become too opaque for humans to supervise effectively. The timing overlaps with separate U.S.-China talks on a proposed notification mechanism for AI incidents reaching national-security significance.20d
    Sam Altman@samaPeople outside the AI labs should have a real say in how this technology develops, and a clear way to judge if it's happening safely. Standards should help prevent the concentration of power, including by making sure new companies and open-model companies can compete. They should also help countries and companies compare evidence and learn from failures. We think the US should lead this effort. Here is our proposal: https://openai.com/index/building-standards-next-phase-ai/19d
    Brian Roemmele@BrianRoemmele@sama I analyzed and responded to Sam’s posting. Unfortunately what he is posting is not what it seems. Give it a read.19d
    Stella Biderman@BlancheMinervaNotably this proposal contains none of: - People outside the labs having a real say it develops - a clear way to judge if it's happening safely. - Standards toprevent concentration of power - Aids to the competitive ability of open-model companies Maybe he posted the wrong one?18d
    Daniel Kokotajlo@DKokotajloRT @trevposts: Per their post today about standards, OpenAI's plan for building AGI is to "build an automated AI researcher" and "iterate w…18d
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet