SlangEstablished
AI Doomer
noun
A label, often dismissive, for someone seen as highly concerned that advanced AI could cause catastrophic harm, potentially including human extinction.
Often a label used in an argument
AI doomer is not a formal technical category.
It is usually a label used in debates about AI risk.
Someone warns that increasingly capable AI systems could become difficult to control. Someone else thinks that concern goes too far.
The AI doomers think every new model is another step toward catastrophe.
Original example written for this dictionary.
The label can refer to people who think AI-driven extinction is a serious possibility.
But it is also used more loosely for people who advocate stronger safeguards, support slowing some forms of development or simply sound pessimistic about advanced AI.
That looseness matters.
Two people called AI doomers may have very different beliefs.
When “doomer” met AI
Doomer existed before the current AI boom as internet slang for someone with a bleak view of the future.
The AI-specific version became much more visible in mainstream technology coverage during 2023 as catastrophic-risk arguments moved into public discussion.
After Geoffrey Hinton left Google and began speaking publicly about AI risks, WIRED published a May 2023 profile titled What Really Made Geoffrey Hinton Into an AI Doomer.
Importantly, that was WIRED's framing.
The article itself shows Hinton taking a more complicated position: he said AI development should continue because it could bring major benefits, while arguing that much more work should go into mitigating possible harms.
That is a useful example of how much information the label can compress — and sometimes hide.
Doomers and accelerationists
Another term helped sharpen the contrast: e/acc, short for effective accelerationism.
The online movement became associated with favoring faster technological development and pushing back against what supporters viewed as excessive pessimism or restraint around AI.
During the turmoil around OpenAI in November 2023, Fortune described Silicon Valley debate in terms of accelerationists and the AI doomer movement, including Marc Andreessen's public embrace of e/acc language.
That created an extremely convenient internet opposition:
accelerationists vs. doomers
It works well as a meme.
It works less well as a precise description of people's beliefs.
Someone can favor rapid AI development while supporting some safety rules.
Someone can take catastrophic risk seriously without supporting a development pause.
People placed in the same “doomer” bucket can disagree sharply with one another.
Where P(doom) comes in
Another piece of vocabulary often appears nearby:
It is shorthand for an estimated probability of an extremely bad future outcome, especially catastrophe associated with advanced AI.
The phrase escaped niche AI circles far enough that it entered U.S. policy discussions.
Axios reported that at a December 2023 Senate AI forum, then-Senate Majority Leader Chuck Schumer opened by asking participants for their P(doom) — described there as the probability that AGI would be disastrous for humanity.
That gives AI doomer an easy stereotype:
someone with a high P(doom).
But the relationship is not exact.
P(doom) is an estimate.
AI doomer is a social label.
Two people could give similar risk estimates and still disagree completely about regulation, research priorities or whether AI development should slow down.
The label also tells you something about the speaker
The word can be descriptive, teasing or openly dismissive.
One person may reserve doomer for people predicting human extinction.
Another may use it for almost anyone arguing that frontier AI deserves more safeguards.
A third may use it jokingly among friends.
So when someone says:
She's an AI doomer.
the sentence may tell you something about the person being described.
It can also tell you how the speaker views AI-risk arguments.
That is why the surrounding conversation matters.
A social label, not a technical category
There is no agreed threshold that makes someone an AI doomer.
People given the label can disagree sharply about both the size of the risk and what should be done about it.
That is what makes the word useful online.
It compresses a complicated disagreement over AI progress, safety, regulation, existential risk and technological optimism into a single recognizable character type.
It also makes the word imprecise.
If you actually want to know what a supposed AI doomer believes, the useful questions are usually more specific:
What risk do they think exists?
How likely do they think it is?
And what do they think should be done about it?