Governance and control

Eliezer Yudkowsky calls for an enforceable global halt to large AI training runs

Eliezer Yudkowsky argued that a six-month pause would not address loss-of-control risk from smarter-than-human AI. He proposed an indefinite worldwide moratorium on large training runs, lower compute ceilings as algorithms improve, GPU tracking, multinational enforcement, and no government or military exemptions.

0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

CURRENT ASSESSMENT · REVISION 1
TOWARD DOOM44confidence 48/100

Why it moved the index

The essay adds a concrete international compute-control mechanism and a specific argument that frontier scaling without solved alignment creates catastrophic loss-of-control risk; magnitude 44 reflects the breadth of the proposed response, while confidence 48 reflects attributable reasoning rather than an observed outcome or independently validated forecast.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 44 · confidence 48

    Initial historical inclusion of Yudkowsky's dated first-person loss-of-control argument and concrete global compute-governance proposal.

    12 Aug 2026