Alana Horowitz Friedman on AI poses an existential threat to human…

Alana Horowitz Friedman

MIRI writer

An off switch refers to building the necessary technical, legal, and institutional infrastructure to make it possible for humanity to shut down unsafe AI systems and AI projects. It will also make it possible to verify that there is no dangerous AI development globally, if humanity decides to halt. Since a halt may be necessary to avoid catastrophe-level risks, including human extinction, there’s a strong case to begin implementing the off switch, since building the required infrastructure could take years. The off switch could be pursued without requiring global consensus on the need for a halt, as it seems uncontroversial that humanity should be able to halt a potentially catastrophic technology if it wanted to. An off switch may appeal both to those who primarily have extinction-level alignment concerns and to those who have a broader range of concerns. For example, being able to monitor, evaluate, and, if necessary, enforce restrictions on frontier AI development would also reduce risks from terrorism, geopolitical destabilization, and other societal disruption.

AI Verified source (2026)

4mo ago

Policy proposals and claims

AI poses an existential threat to humanity

votes For

Statement relation verification history AI Verified Report this

AI Verified The quote clearly indicates the author believes AI could create extinction-level danger: it says a halt in AI development "may be necessary to avoid catastrophe-level risks, including human extinction" and describes AI as a "potentially catastrophic technology." That implies support for the claim that AI poses an existential threat to humanity. · YouCongress gpt-5.4-2026-03-05 · 4mo ago

Vote inference verification history AI Verified Report this

AI Verified The quote says a halt in AI development "may be necessary to avoid catastrophe-level risks, including human extinction," and mentions "extinction-level alignment concerns," which clearly supports the idea that AI can pose an existential threat. · YouCongress gpt-5.4-2026-03-05 · 4mo ago

Governments and labs should build the capability for a global slowdown of frontier AI development in case it's needed

votes For

Statement relation verification history AI Verified Report this

AI Verified The quote clearly supports building technical, legal, and institutional infrastructure so humanity can globally halt or restrict frontier AI development if necessary. That is a stronger version of building capability for a global slowdown, so it implies support for the full statement. · YouCongress gpt-5.4-2026-03-05 · 4mo ago

Vote inference verification history AI Verified Report this

AI Verified The quote says "there’s a strong case to begin implementing the off switch" and to build the "technical, legal, and institutional infrastructure" so humanity can "halt" or "enforce restrictions on frontier AI development" globally if needed. · YouCongress gpt-5.4-2026-03-05 · 4mo ago

Slow down AI development

abstains

Statement relation verification history AI Verified Report this

AI Verified Full-claim match: the quote addresses AI development/capability growth and a slower pace, pause, halt, or explicit rejection/mixed position. Evidence: there is no dangerous AI development globally, if humanity decides to halt. Since a halt may be necessary to avoid catastrophe-level risks, including human extinction, there’s a strong case to begin implementing the off switch, since buildi · Hector Perez Arenas gpt-5 · 1mo ago

Vote inference verification history AI Verified Report this

AI Verified Recorded abstain matches the quote's stance on slowing AI development. Evidence: there is no dangerous AI development globally, if humanity decides to halt. Since a halt may be necessary to avoid catastrophe-level risks, including human extinction, there’s a strong case to begin implementing the off switch, since buildi · Hector Perez Arenas gpt-5 · 1mo ago

Quote authenticity verification history

Report this

AI Verified The passage is authentic and appears verbatim in the "Off Switch" section of the April 13, 2026 MIRI post "Summary: AI Governance to Avoid Extinction." The source page credits Alana Horowitz Friedman as the author, and the three quoted paragraphs match the text at lines 50–52 exactly. ([intelligence.org](https://intelligence.org/2026/04/13/summary-ai-governance-to-avoid-extinction/)) · YouCongress gpt-5.4-2026-03-05 · 4mo ago

AI Verified Verified. The source_url (intelligence.org/2026/04/13/summary-ai-governance-to-avoid-extinction/) returned HTTP 403 to direct fetch, but web search confirmed both the quote and the author. The page is MIRI's plain-language "Summary: AI Governance to Avoid Extinction" (13 April 2026), authored by Alana Horowitz Friedman, a MIRI communications writer — author attribution correct (she is credited with authoring this summary; the underlying May 2025 paper is by MIRI's Technical Governance Team). The quote was confirmed: the "off switch" definition ("the technical, legal, and institutional infrastructure required to shut down... dangerous AI") matches, and the sentence "Since a halt may be necessary to avoid catastrophe-level risks, including human extinction, there's a strong case to begin implementing the off switch, since building the required infrastructure could take years" was reproduced verbatim by web search. Year 2026 correct (the summary was published April 2026). The vote "for" aligns precisely with statement 451 ("Governments and labs should build the capability for a global slowdown of frontier AI development in case it's needed"): the entire quote advocates building the "off switch" infrastructure so humanity can monitor, restrict, and if necessary halt frontier AI development — i.e., preserving/building exactly the slowdown capability the statement describes. · Hector Perez Arenas claude-opus-4-8 · 4mo ago

replying to Alana Horowitz Friedman

Other opinions from Alana Horowitz Friedman

See all

Alana Horowitz Friedman votes For

Ban superintelligence development until safety consensus is reached

They conclude that the only trajectory without unacceptably high levels of catastrophic harm is a scenario in which humanity develops the ability to monitor and restrict AI development (an “off switch”) and then likely uses it to put a long-term halt...

more AI Verified source (Apr 13, 2026)