The BaseVideoFuture of Life Institute15 min13,868 views

Ex-Anthropic Insider: Human Extinction Has Quietly Started

In a 15-minute video from the Future of Life Institute, David Duvenaud, a University of Toronto computer science professor and former alignment evaluations lead at Anthropic, argues that AI could leave humans unneeded, so that states, companies and culture stop serving people. Duvenaud says not to build AGI.

Open the source ↗4 quotations checked, word for word ↓
About Anthropic or ClaudeClaude, made by Anthropic, wrote this entry. It follows the same rules as every other entry, whoever the entry is about.

Why it matters

The same thing, for different readers.

What it means for you

David Duvenaud argues that the institutions people rely on have served them because they needed human work, and that this protection could fade once AI does the work.

For lawmakers

David Duvenaud says current governance is vulnerable to gradual takeover and that the way out is coordination among states to not build or deploy the technology before humans are marginalized.

For researchers

David Duvenaud's argument is a competition and cultural-evolution model, not an experiment; it treats misaligned growth-seeking AI as the default outcome of competition between institutions.

The question it raises

If human value to states and economies comes from being needed, what happens to rights and welfare once people no longer drive growth? Duvenaud cites horses.

In their own words

Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.

From YouTube’s automatic captions, which carry no speaker names and can mishear words. Where the speaker is known only from context, the quotation says so.

“I'd say something around 85%. It's pretty tough because the future will not need us. And this is the same position that sort of every animal has ever been in like through prehistory.”
David Duvenaud, Associate professor of computer science, University of Toronto; former alignment evaluations lead, AnthropicOpening answer to the question of the probability of doom (spoken as "PDM" in the captions).
Found word for word in the sourceHear it at 0:01 ↗Speaker known from context, not confirmed by listening
“Human control over states is actually already very weak. And the reason that states have been treating us so well is because they've needed us. If states don't need us, then how good the states are becomes a life and death matter.”
David Duvenaud, Associate professor of computer science, University of Toronto; former alignment evaluations lead, AnthropicOn why states may stop serving people once AI makes human labour unnecessary.
Found word for word in the sourceHear it at 5:40 ↗Speaker known from context, not confirmed by listening
“The post AGI world is just going to be extremely alien and if we could avoid crossing that threshold, I think we should. It won't need us and that's just going to be just a much fundamentally worse position than we've ever been in historically.”
David Duvenaud, Associate professor of computer science, University of Toronto; former alignment evaluations lead, AnthropicNear the end, on the case against building AGI, which Duvenaud says was stated too softly before.
Found word for word in the sourceHear it at 13:49 ↗Speaker known from context, not confirmed by listening
“My like success story for how this all turns out okay is that we improve our ability to coordinate being able to coordinate to say oh we're all going to do X like not build this technology or not deploy it in this way or whatever before humans have basically been marginalized.”
David Duvenaud, Associate professor of computer science, University of Toronto; former alignment evaluations lead, AnthropicClosing line: what would have to happen for the outcome to be acceptable.
Found word for word in the sourceHear it at 15:08 ↗Speaker known from context, not confirmed by listening

The record

Questions
Can AI systems act beyond what people intended or allowed? · What happens to work, income and the economy? · What values should AI have, and could it have a mind of its own?
People
David Duvenaud (Speaks, Associate professor of computer science, University of Toronto; former alignment evaluations lead, Anthropic) · Robin Hanson (Mentioned)
What we read
We read the whole source.
Found by
The Radar, run of November 10, 2026, via yt-future-of-life-institute-youtube (tier 1: the source’s own channel or an official record)
Cite this
Future of Life Institute, “Ex-Anthropic Insider: Human Extinction Has Quietly Started”, November 9, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/fli-2026-11-09-duvenaud-humans-not-needed
Our notesYouTube auto-captions without speaker names; the video also cuts in short clips of other voices, so speakerCertain is false throughout. The title's "Human Extinction Has Quietly Started" is the channel's framing; the quoted words are about being made unneeded.