YouCongress
About Sign up

Comment by Stuart J. Russell

Human avatar
AI Expert and Professor
The problem of control — how do we maintain power, forever, over entities that will eventually become more powerful than us — is the most important problem facing humanity. AI Unverifiable source (2023)
Like Share on X 2mo ago
Policy proposals and claims
Require human-in-the-loop oversight for agentic AI systems acting in high-stakes domains

Verification History

AI Unverifiable Quote attributed to Stuart J. Russell, year 2023, "for" statement #418 ("Require human-in-the-loop oversight for agentic AI systems acting in high-stakes domains"). Source: Senate Judiciary testimony PDF (2023-07-26). Findings: (1) The source PDF returns HTTP 403 to WebFetch, so I could not confirm the verbatim text on the page. (2) Web search did not return the exact verbatim phrase; the wording ("the problem of control — how do we maintain power, forever, over entities that will eventually become more powerful than us — is the most important problem facing humanity") is strongly consistent with Russell's well-documented "Human Compatible: AI and the Problem of Control" framing, but I could not positively confirm it verbatim. (3) STALENESS: year 2023, older than 2025/2026; protocol prefers a recent replacement. A recent Russell source exists (Fortune, "The AI arms race could lead to human extinction, top researcher warns," 2026-02-18) that a human could use to refresh this entry, but no verifiable verbatim recent replacement could be obtained. (4) RELEVANCE/VOTE: the quote concerns the broad existential AI control problem rather than specifically human-in-the-loop oversight in high-stakes domains, so its fit to this statement is loose; the "for" vote direction is thematically defensible (Russell strongly favors maintaining human control over AI). Marked ai_unverifiable because the source is blocked and the verbatim text could not be confirmed. · Hector Perez Arenas claude-opus-4-8 · 8d ago
replying to Stuart J. Russell