We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Comment by Jules Palayer
SIPRI researcher on AI and security
Finally, to avoid that open-source models are accessed and retrofitted for malicious purposes a potential solution is to create self-destruct codes that activate if the model is tampered with.Disputed source (Mar 26, 2024)
9mo ago
Policy proposals and claims
abstains
No statement relation verification comments yet.
No vote answer verification comments yet.
Quote authenticity verification history
Report thisQuote authenticity comments
Disputed
The wording is real: an archived UNODA blog page contains the sentence verbatim, continuing after your excerpt with "that activate if the model is tampered with." ([archive-disarmament.unoda.org](https://archive-disarmament.unoda.org/responsible-innovation-ai/blog/)) But it is not correctly attributed to the United Nations Office for Disarmament Affairs as the author. The quote appears in the UNODA-hosted post "Unpacking the concerns around AI and biotechnology" (26 March 2024), and the blog credits that post to Jules Palayer; the URL you gave is the general blog index, not the specific article page. ([archive-disarmament.unoda.org](https://archive-disarmament.unoda.org/responsible-innovation-ai/blog/))
·
YouCongress
gpt-5.4-2026-03-05
· 19d ago
AI Unverifiable
Could not verify the exact quote. The source URL (disarmament.unoda.org/en/responsible-innovation-ai/blog) returned 403 Forbidden when fetched directly. Multiple targeted web searches for the phrase 'self-destruct codes' in connection with UNODA, open-source AI models and retrofitting did not return matching content - search engines surfaced UNODA/SIPRI's responsible innovation initiative and academic 'self-destructing models' research, but not this specific UNODA blog passage. The URL points to a generic blog index rather than a specific post, no year is provided, and the quote is mid-sentence ('Finally, to avoid that...'). Self-destruct mechanisms for AI models is a real academic concept (e.g., Henderson et al. 2022), so the quote is plausible but not confirmed. Vote 'abstain' is reasonable for UNODA, which generally surveys options rather than advocating bans.
·
Hector Perez Arenas
claude-opus-4-7
· 1mo ago
replying to Jules Palayer