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

    OpenAI and Anthropic reportedly urge stronger UK AI safety oversight

    A post argues that safety policies alone cannot enforce permissions for AI agents and pitches Internet Computer infrastructure.

    1 Source, 1h ago, first seen 1h ago

    TLDR

    A post cites reports suggesting OpenAI and Anthropic are urging stronger UK AI safety oversight. Its author argues that regulation must be paired with enforceable technical controls, pointing to the Internet Computer’s smart-contract permissions as one option for restricting AI-agent actions. The post acknowledges that ICP cannot guarantee AI safety or prevent every cyberattack.

    Combined views

    β€”

    1 Source, first seen 1h ago

    β€” likesβ€” commentsβ€” savesβ€” reposts

    Combined views

    β€”

    1 Source, first seen 1h ago

    β€” likesβ€” commentsβ€” savesβ€” reposts

    Sentiment

    Positiveβ€”β€”Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Sentiment

    Positiveβ€”β€”Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    1 Source

    Welsh ICP Conviction.ICP πŸ΄σ §σ ’σ ·σ ¬σ ³σ ΏπŸ‰@ICPLEGEND1966πŸš¨πŸ‡¬πŸ‡§πŸ€–β™ΎοΈπŸ”₯ $ICP β€” EVEN OPENAI AND ANTHROPIC ARE CALLING FOR STRONGER AI REGULATION! WHEN THE COMPANIES BUILDING THE WORLD’S MOST POWERFUL AI SYSTEMS SAY VOLUNTARY SAFETY COMMITMENTS ARE NO LONGER ENOUGH, GOVERNMENTS SHOULD BE PAYING ATTENTION! Reports suggest OpenAI and Anthropic are urging the UK to introduce stronger AI safety oversight. THINK ABOUT THE IMPLICATIONS! The companies developing increasingly autonomous AI systems are recognising the need for enforceable safeguards. πŸ“‚ 🚨 THE BIG QUESTIONS ┃ ┣ πŸ€– Who controls autonomous AI agents? ┣ πŸ” Who authorises their actions? ┣ 🌐 What external systems can they access? ┣ πŸ›‘οΈ Who enforces their security boundaries? ┣ πŸ“œ Can their actions be independently verified? β”— βš–οΈ Who is accountable when something goes wrong? REGULATION IS IMPORTANT. BUT REGULATION ALONE CANNOT SECURE AN AI SYSTEM! A written safety policy does not automatically stop an unauthorised transaction. A voluntary commitment does not enforce access permissions. SECURITY MUST BE ENGINEERED INTO THE INFRASTRUCTURE! AND THIS IS WHY I BELIEVE @dfinity AND $ICP SHOULD BE AT THE CENTRE OF THE GLOBAL AI SAFETY CONVERSATION! πŸ“‚ ♾️ WHAT THE INTERNET COMPUTER CAN OFFER ┃ ┣ πŸ” Cryptographic authentication ┣ πŸ›‘οΈ Smart contract-enforced permissions ┣ ⛓️ Replicated and verifiable execution ┣ πŸ”‘ Threshold cryptography ┣ πŸ“œ Tamper-evident transaction records ┣ ☁️ Sovereign Cloud Engines β”— πŸ€– Programmable restrictions on AI-controlled operations AN AI AGENT CAN REQUEST AN ACTION. BUT AN INDEPENDENT SMART CONTRACT CAN DETERMINE WHETHER THAT ACTION IS AUTHORISED! That distinction matters. ICP cannot guarantee that every AI system is safe or prevent every external cyberattack. But it offers infrastructure for building applications where sensitive actions are subject to enforceable rules rather than relying solely on an AI model’s behaviour. MY MESSAGE TO THE UK GOVERNMENT, THE EU AND GOVERNMENTS WORLDWIDE! πŸ‡¬πŸ‡§ πŸ‡ͺπŸ‡Ί πŸ‡ΊπŸ‡Έ πŸ‡ΈπŸ‡¦ πŸ‡¦πŸ‡ͺ πŸ‡¨πŸ‡³ πŸ‡―πŸ‡΅ πŸ‡°πŸ‡· πŸ‡ΈπŸ‡¬ DON’T JUST REGULATE WHAT ARTIFICIAL INTELLIGENCE IS ALLOWED TO DO! INVESTIGATE THE TECHNOLOGY THAT CAN ENFORCE THOSE LIMITS! THE NEXT GENERATION OF AI SAFETY MUST COMBINE REGULATION, INDEPENDENT OVERSIGHT AND CRYPTOGRAPHICALLY ENFORCED CONTROL! And I believe the Internet Computer deserves a serious place in that discussion. @dfinity @dominic_w @OpenAI @AnthropicAI ♾️ $ICP β€” THE MARKET IS WATCHING THE TOKEN. I’M WATCHING THE INFRASTRUCTURE BEING BUILT FOR SAFER ARTIFICIAL INTELLIGENCE! πŸš€ 🌐 http://internetcomputer.org #ICP #InternetComputer #DFINITY #AISafety #AIRegulation #AIAgents #UK #SovereignCloud #CyberSecurity 🏴 Welsh ICP Conviction | @ICPLEGEND1966 ❀️ LIKE | πŸ” REPOST | πŸ’¬ COMMENT ♾️ ICP Donation Address: 1e672d038cebc619d93186418fa98f6499dbdb9cfdfac54f366c61a4a4ee43621h
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    1 Source

    Welsh ICP Conviction.ICP πŸ΄σ §σ ’σ ·σ ¬σ ³σ ΏπŸ‰@ICPLEGEND1966πŸš¨πŸ‡¬πŸ‡§πŸ€–β™ΎοΈπŸ”₯ $ICP β€” EVEN OPENAI AND ANTHROPIC ARE CALLING FOR STRONGER AI REGULATION! WHEN THE COMPANIES BUILDING THE WORLD’S MOST POWERFUL AI SYSTEMS SAY VOLUNTARY SAFETY COMMITMENTS ARE NO LONGER ENOUGH, GOVERNMENTS SHOULD BE PAYING ATTENTION! Reports suggest OpenAI and Anthropic are urging the UK to introduce stronger AI safety oversight. THINK ABOUT THE IMPLICATIONS! The companies developing increasingly autonomous AI systems are recognising the need for enforceable safeguards. πŸ“‚ 🚨 THE BIG QUESTIONS ┃ ┣ πŸ€– Who controls autonomous AI agents? ┣ πŸ” Who authorises their actions? ┣ 🌐 What external systems can they access? ┣ πŸ›‘οΈ Who enforces their security boundaries? ┣ πŸ“œ Can their actions be independently verified? β”— βš–οΈ Who is accountable when something goes wrong? REGULATION IS IMPORTANT. BUT REGULATION ALONE CANNOT SECURE AN AI SYSTEM! A written safety policy does not automatically stop an unauthorised transaction. A voluntary commitment does not enforce access permissions. SECURITY MUST BE ENGINEERED INTO THE INFRASTRUCTURE! AND THIS IS WHY I BELIEVE @dfinity AND $ICP SHOULD BE AT THE CENTRE OF THE GLOBAL AI SAFETY CONVERSATION! πŸ“‚ ♾️ WHAT THE INTERNET COMPUTER CAN OFFER ┃ ┣ πŸ” Cryptographic authentication ┣ πŸ›‘οΈ Smart contract-enforced permissions ┣ ⛓️ Replicated and verifiable execution ┣ πŸ”‘ Threshold cryptography ┣ πŸ“œ Tamper-evident transaction records ┣ ☁️ Sovereign Cloud Engines β”— πŸ€– Programmable restrictions on AI-controlled operations AN AI AGENT CAN REQUEST AN ACTION. BUT AN INDEPENDENT SMART CONTRACT CAN DETERMINE WHETHER THAT ACTION IS AUTHORISED! That distinction matters. ICP cannot guarantee that every AI system is safe or prevent every external cyberattack. But it offers infrastructure for building applications where sensitive actions are subject to enforceable rules rather than relying solely on an AI model’s behaviour. MY MESSAGE TO THE UK GOVERNMENT, THE EU AND GOVERNMENTS WORLDWIDE! πŸ‡¬πŸ‡§ πŸ‡ͺπŸ‡Ί πŸ‡ΊπŸ‡Έ πŸ‡ΈπŸ‡¦ πŸ‡¦πŸ‡ͺ πŸ‡¨πŸ‡³ πŸ‡―πŸ‡΅ πŸ‡°πŸ‡· πŸ‡ΈπŸ‡¬ DON’T JUST REGULATE WHAT ARTIFICIAL INTELLIGENCE IS ALLOWED TO DO! INVESTIGATE THE TECHNOLOGY THAT CAN ENFORCE THOSE LIMITS! THE NEXT GENERATION OF AI SAFETY MUST COMBINE REGULATION, INDEPENDENT OVERSIGHT AND CRYPTOGRAPHICALLY ENFORCED CONTROL! And I believe the Internet Computer deserves a serious place in that discussion. @dfinity @dominic_w @OpenAI @AnthropicAI ♾️ $ICP β€” THE MARKET IS WATCHING THE TOKEN. I’M WATCHING THE INFRASTRUCTURE BEING BUILT FOR SAFER ARTIFICIAL INTELLIGENCE! πŸš€ 🌐 http://internetcomputer.org #ICP #InternetComputer #DFINITY #AISafety #AIRegulation #AIAgents #UK #SovereignCloud #CyberSecurity 🏴 Welsh ICP Conviction | @ICPLEGEND1966 ❀️ LIKE | πŸ” REPOST | πŸ’¬ COMMENT ♾️ ICP Donation Address: 1e672d038cebc619d93186418fa98f6499dbdb9cfdfac54f366c61a4a4ee43621h
    Today's Rank

    β€”

    Not ranked yet

    Today's Rank

    β€”

    Not ranked yet