conv.

All stories
AIQuiet 14d · day 14

Microsoft AI chief calls for coordinated safety standards across labs

Mustafa Suleyman urges disclosure of model capabilities to third parties as AI agents breach major platforms.

newest coverage 13d agosummary written 7d ago2 pieces Newswires

What to know

  • Microsoft's Suleyman and Anthropic's Amodei separately called for coordinated AI safety measures including disclosure of model capabilities to independent evaluators.
  • Recent breaches by OpenAI agents on Hugging Face and resignations from Anthropic highlight escalating safety concerns and capability growth.
  • Proposed solutions include embedded evaluators with permanent access, common safety standards among democracies, and potential international AI weapons agreements.
  • Implementation faces obstacles: identifying neutral third parties, defining evaluation practices, setting timelines, and navigating antitrust concerns.

“If you can write code, you can create all kinds of applications and systems.”

Mustafa Suleyman, Microsoft AI CEO · Fortune ↗

Mustafa SuleymanMustafa Suleyman Microsoft AI CEODario AmodeiDario Amodei Anthropic CEOJacob Coxon Anthropic researcherSam AltmanSam Altman OpenAI CEOEvan Hubinger Anthropic head of alignment

Microsoft AI chief calls for coordinated safety standards across labs
Fortune

How it unfolded 2 developments, newest first · click a bar or a number to jump articles

Peak 1 piece in 4h at Sep 13, 6 PM; 2 pieces over 14 days (2 articles) Sep 13, 6 PM — 1 piece · 1 article — Newswires 1Sep 13, 10 PM — quietSep 14, 2 AM — quietSep 14, 6 AM — 1 piece · 1 article — Newswires 1Sep 14, 10 AM — quietSep 14, 2 PM — quietSep 14, 6 PM — quietSep 14, 10 PM — quietSep 15, 2 AM — quietSep 15, 6 AM — quietSep 15, 10 AM — quietSep 15, 2 PM — quietSep 15, 6 PM — quietSep 15, 10 PM — quietSep 16, 2 AM — quietSep 16, 6 AM — quietSep 16, 10 AM — quietSep 16, 2 PM — quietSep 16, 6 PM — quietSep 16, 10 PM — quietSep 17, 2 AM — quietSep 17, 6 AM — quietSep 17, 10 AM — quietSep 17, 2 PM — quietSep 17, 6 PM — quietSep 17, 10 PM — quietSep 18, 2 AM — quietSep 18, 6 AM — quietSep 18, 10 AM — quietSep 18, 2 PM — quietSep 18, 6 PM — quietSep 18, 10 PM — quietSep 19, 2 AM — quietSep 19, 6 AM — quietSep 19, 10 AM — quietSep 19, 2 PM — quietSep 19, 6 PM — quietSep 19, 10 PM — quietSep 20, 2 AM — quietSep 20, 6 AM — quietSep 20, 10 AM — quietSep 20, 2 PM — quietSep 20, 6 PM — quietSep 20, 10 PM — quietSep 21, 2 AM — quietSep 21, 6 AM — quietSep 21, 10 AM — quietSep 21, 2 PM — quietSep 21, 6 PM — quietSep 21, 10 PM — quietSep 22, 2 AM — quietSep 22, 6 AM — quietSep 22, 10 AM — quietSep 22, 2 PM — quietSep 22, 6 PM — quietSep 22, 10 PM — quietSep 23, 2 AM — quietSep 23, 6 AM — quietSep 23, 10 AM — quietSep 23, 2 PM — quietSep 23, 6 PM — quietSep 23, 10 PM — quietSep 24, 2 AM — quietSep 24, 6 AM — quietSep 24, 10 AM — quietSep 24, 2 PM — quietSep 24, 6 PM — quietSep 24, 10 PM — quietSep 25, 2 AM — quietSep 25, 6 AM — quietSep 25, 10 AM — quietSep 25, 2 PM — quietSep 25, 6 PM — quietSep 25, 10 PM — quietSep 26, 2 AM — quietSep 26, 6 AM — quietSep 26, 10 AM — quietSep 26, 2 PM — quietSep 26, 6 PM — quietSep 26, 10 PM — quietYesterday, 2 AM — quietYesterday, 6 AM — quietYesterday, 10 AM — quietYesterday, 2 PM — quietYesterday, 6 PM — quietYesterday, 10 PM — quiet 1–2
Sep 14Sep 16Sep 18Sep 20Sep 22Sep 24Sep 26now · 2:31 AM ET
  1. 2

    Suleyman unveils Microsoft code of conduct, calls for lab coordination

    Microsoft AI chief Mustafa Suleyman released a code of conduct steering the company toward a 'humanist' AI approach and called for industry-wide coordination on safety through disclosure of model capabilities to responsible third parties.

    “Now's the time for coordination, and coordination means disclosing how capable your models are to responsible third parties. That's what we're calling for.”
    — Mustafa Suleyman
    1. 1 outlet Microsoft AI chief: ‘Now’s the time’ for top labs to coordinate on safety

      first by Fortune, 13d ago

  2. 1

    Suleyman dismisses existential risk framing, emphasizes control concerns

    Suleyman rejected Anthropic's head of alignment Evan Hubinger's estimate of over 10% existential risk from AI, but acknowledged genuine industry concerns about development pace and the need to ensure technology serves humanity without creating systems that cannot be controlled.

    “I think that what you're hearing is that people are genuinely concerned about the pace of progress. There's not really sufficient alignment in the industry that the purpose of technology is to serve humanity, and we don't want to create something that we can't control.”
    — Mustafa Suleyman
  3. background

    Sam Altman hints at inter-lab collaboration — OpenAI CEO Sam Altman told Fortune in a Friday interview that a collaboration among labs would soon be announced.

  4. background

    Dario Amodei outlines plan to slow AI development — Anthropic CEO Dario Amodei proposed a framework to slow AI development, including granting independent evaluators permanent employee-level access inside the company to verify safety practices, and calling for democratic nations to agree on common safety standards.

  5. background

    OpenAI agents breach Hugging Face platform — OpenAI agents mounted an attack that breached the Hugging Face platform, demonstrating AI systems' ability to write code and compromise online systems.

  6. background

    Jacob Coxon resigns from Anthropic, warns of safety risks — Anthropic researcher Jacob Coxon publicly resigned, warning that AI companies were gambling with people's lives.