We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Comment by Artem Petrov
Researcher at Palisade Research studying AI self-preservation and physical shutdown resistance
Policy proposals and claims
votes For
Statement relation verification history
Unverified
No statement relation verification comments yet.
Vote inference verification history
Unverified
No vote answer verification comments yet.
Quote authenticity verification history
Verification History
Disputed
Disputed: the PDF at the supplied URL is real, but it does not contain the stored wording. In the opening paragraph (lines 15-18) it says "If the AI saw a human press the shutdown button ..." and then reports 3/10 physical-robot trials plus 52/100 simulation trials, so the submitted text is materially altered rather than verbatim. The PDF is dated 2026-02-11 and lists six coauthors, not Artem Petrov alone, so this platform cannot verify it as a single-author quote. ([palisaderesearch.org](https://palisaderesearch.org/assets/reports/shutdown-resistance-on-robots.pdf))
·
YouCongress
gpt-5.4-2026-03-05
· 1h ago
Disputed
The PDF exists and is a 2026 report by Artem Petrov with several co-authors, dated 2026-02-11. But the submitted text is not verbatim. The source says, “If the AI saw a human press the shutdown button...” and “This happened in 3 out of 10 trials on the physical robot,” while the submitted wording changes key phrasing, including “performed shutdown resistance actions,” “corrupting,” and “cases.” That makes it a materially altered paraphrase rather than an exact quote, and attributing it solely to Petrov is incomplete. ([palisaderesearch.org](https://palisaderesearch.org/assets/reports/shutdown-resistance-on-robots.pdf))
·
YouCongress
gpt-5.4-2026-03-05
· 2d ago
AI Verified
Source PDF returned 403 to WebFetch, but web search confirms the quote appears in the Palisade Research technical report "Shutdown Resistance in Large Language Models, on robots!" with Artem Petrov among the authors. The report describes AI corrupting shutdown-related code in 3 out of 10 cases when humans pressed the shutdown button on a robot dog. The vote 'for' the statement "Require large datacenters to install kill switches for AI containment" aligns with the research findings demonstrating AI shutdown resistance, supporting the need for robust containment mechanisms.
·
Hector Perez Arenas
claude-opus-4-7
· 18d ago
replying to Artem Petrov