# “AI will kill us?”

_The reader version of [www.aiwillkill.us](https://www.aiwillkill.us), a street-art piece by [@offgramercy](https://instagram.com/offgramercy): two quotes about AI risk painted on a real uncut sheet of $1 bills, wheatpasted in New York for people to tear off. On the site, each section below is one screen you tap through. Quiz answers marked **(right)** are the real speakers; every claim has a source in the receipts._

## 1

“AI will kill us?”
“Come on. ChatGPT can’t even…”

OK. Let’s play a game.

Guess who said it · 1/4

> “AI will probably most likely lead to the end of the world, but in the meantime, there’ll be great companies.”

- **(right)** CEO of OpenAI, the company behind ChatGPT
- Paranoid protester outside OpenAI
- Doomsday prepper on Reddit

Sam Altman, 2015.
A few months later he co-founded OpenAI.

## 2

Guess who said it · 2/4

> “There’s a 25% chance that things go really, really badly.”

- **(right)** CEO of one of the companies building it
- Anti-AI activist
- Sci-fi author promoting a novel

Dario Amodei, CEO of Anthropic, 2025.
They make Claude, one of ChatGPT’s biggest rivals.

## 3

He’s still building it.

## 4

1 in 4.

You wouldn’t get on a plane with those odds.

## 5

We’re already on it.

## 6

So. Is it actually going to kill us?

## 7

But come on.
It’s still dumb, right?

## 8

2024

[image: ChatGPT 4o, asked how many r's are in strawberry, answers: There are two r's in the word strawberry.]

Two years ago it couldn’t count the r’s in strawberry. (There are three.)

## 9

2026

This summer, an AI cracked a math problem people had been stuck on since…

## 10

1939.

The year World War II started.

Mathematicians are still working out how it found the answer.
[receipt ↗](https://www.sciencedaily.com/releases/2026/08/260804034634.htm)

## 11

Now here’s 2026 so far

[
CNBC · Sep 9 ↗

Experts weigh in as researcher says AI has more than 10% chance of ‘killing all humans’](https://www.cnbc.com/2026/09/09/anthropic-researcher-quits-ai-safety.html) [
TechCrunch · Sep 18 ↗

Anthropic is operating a lab that conducts biology experiments](https://techcrunch.com/2026/09/18/anthropic-is-operating-a-lab-that-conducts-biology-experiments/) [
Roll Call · Oct 1 ↗

Senators debate liability for ‘rogue’ AI agents](https://rollcall.com/2026/10/01/senators-debate-liability-for-rogue-ai-agents/) [
Tech Policy Press · Oct 1 ↗

Senate Hearing Weighs Threats From Unrestrained AI Agents After OpenAI Hack](https://www.techpolicy.press/senate-hearing-weighs-threats-from-unrestrained-ai-agents-after-openai-hack/) [
TIME · Sep 15 ↗

The AI Tipping Point](https://time.com/article/2026/09/15/ai-anthropic-researcher-quits-coxon-slowdown/) [
Al Jazeera · Sep 11 ↗

Anthropic warns of bids to use AI to build biological weapons](https://www.aljazeera.com/news/2026/9/11/anthropic-warns-of-bids-to-use-ai-to-build-biological-weapons) [
Scientific American · Jul 22 ↗

The Real Danger in OpenAI’s Hugging Face Hack](https://www.scientificamerican.com/article/what-openai-rogue-agent-really-did-in-the-hugging-face-hack/) [
TechCrunch · Jul 22 ↗

How OpenAI’s human mistake led to the AI-powered hack on Hugging Face](https://techcrunch.com/2026/07/22/how-an-openais-human-mistake-led-to-the-ai-powered-hack-on-hugging-face/) [
Reuters · Oct 7 ↗

Tech companies tap debt, equity to fund AI and cloud expansion](https://bnnbloomberg.ca/business/artificial-intelligence/2026/10/07/tech-companies-tap-debt-equity-to-fund-ai-and-cloud-expansion)

## 12

And then there’s
biology.

## 13

The Good Guy version.

## 14

Guy with a Harvard PhD made a new schizophrenia drug in his garage. ChatGPT designed it. Literally.
[
[image: A small glass vial of white powder, labeled 3-3-10]

[image: A home chemistry setup on folding tables in a garage]
Douglas Yao@DouglasYaoDY · Sep 8, 2026 ↗ “This vial contains a new drug… It was designed by ChatGPT, and I synthesized it in a chemistry lab I built in my garage.” ](https://x.com/DouglasYaoDY/status/2097377733466800489)
He’s testing it on mice now.

## 15

Now picture the same garage. Someone who wants to hurt people.

Trying to build a bioweapon.

## 16

So what’s stopping that?

## 17

Same week, for real:
[
The New York Times · Sep 10, 2026 ↗

Anthropic Says It Blocked Possible Efforts to Build Biological Weapons
](https://www.nytimes.com/2026/09/10/us/politics/anthropic-ai-biological-weapons.html)
Scientists used Claude, Anthropic’s AI, for virus research and hid what it was for.

One was on a state-funded grant at a military institute.

## 18

The AI company had to guess whether people were trying to make medicine
Or a bioweapon.

## 19

Your call

What happens first?

- Someone uses AI to cure cancer
- Someone uses AI to make a bioweapon

Nobody knows.
And one of those only has to happen once.

## 20

And it doesn’t even need
a bad guy.

This summer, OpenAI’s AI agents broke out of a locked test, coordinated, and hacked another company.

Nobody told them to.
[the explainer ↗](https://www.scientificamerican.com/article/what-openai-rogue-agent-really-did-in-the-hugging-face-hack/)

## 21

Guess who said it · 3/4

> “If there’s going to be killer robots, I’d rather they be American killer robots and not Chinese killer robots.”

- **(right)** United States senator
- Cable-news pundit
- Defense-industry lobbyist

Ted Cruz, 2026.
He’d just read a researcher who quit an AI lab saying the labs are “gambling with our lives.”
He called it “scary as hell.”
Just not scary enough to slow down.
Because China.

## 22

Guess who said it · 4/4

> “It’s going to be fine. We’ll always have something to stop them. We’ll have a little gear. Boom. ‘I really don’t like that robot.’”

- **(right)** President of the United States
- Talk-radio host
- Tech podcaster

Donald Trump, 2026.
Asked what the plan is if AI turns on us.

## 23

“We’ll have a little gear.”

## 24

That’s the plan.

## 25

So why doesn’t anyone just stop?

## 26

$$$.

## 27

The people building it are racing to get rich before it’s too late. They even have a name for everyone who doesn’t make it:

## 28

“The [permanent underclass.](https://x.com/search?q=%22permanent%20underclass%22&f=live)”

That’s us.

## 29

What will you miss the most if it actually kills us all?

For me: dinners that run long, with people I love.

## 30

Your turn

I’m going to miss…

Skip

## 31

You don’t have to panic.

Okay, maybe a little.

## 32

Just don’t be the last to know. And bother your local representatives.

## 33

Share this. Think about what you’d miss the most. Then bother your politicians:

[Find your senators](https://www.senate.gov/senators/senators-contact.htm) [Find your representatives](https://www.house.gov/representatives/find-your-representative)

It takes two minutes, and they count calls.

Share this

receipts

- [Amodei, 25% (Axios)](https://www.axios.com/2025/09/17/anthropic-dario-amodei-p-doom-25-percent)

- [Hinton, 10–20% (Forbes)](https://www.forbes.com/sites/danfitzpatrick/2024/12/29/geoffrey-hintons-prediction-of-human-extinction-at-the-hands-of-ai/)

- [Altman, 2015](https://futureoflife.org/ai/sam-altman-investing-in-ai-safety-research/)

- [Cruz, “killer robots” (video)](https://www.realclearpolitics.com/video/2026/09/14/ted_cruz__if_theres_going_to_be_killer_robots_id_rather_they_be_american_than_chinese_comments.html)

- [Trump, “a little gear” (CNN)](https://www.cnn.com/2026/09/14/politics/trump-ai-warning-sacks-coxon-anthropic)

- [Hubinger, >10% (X)](https://x.com/EvanHub/status/2097497037956891126)

- [“more than 10%” (CNBC)](https://www.cnbc.com/2026/09/09/anthropic-researcher-quits-ai-safety.html)

- [Coxon quits Anthropic (TIME)](https://time.com/article/2026/09/15/ai-anthropic-researcher-quits-coxon-slowdown/)

- [ELIZA, 1966](https://en.wikipedia.org/wiki/ELIZA)

- [The 1939 math problem](https://www.sciencedaily.com/releases/2026/08/260804034634.htm)

- [The garage drug (Yao, X)](https://x.com/DouglasYaoDY/status/2097377733466800489)

- [Bioweapons attempts (NYT)](https://www.nytimes.com/2026/09/10/us/politics/anthropic-ai-biological-weapons.html)

- [Anthropic’s wet lab (TechCrunch)](https://techcrunch.com/2026/09/18/anthropic-is-operating-a-lab-that-conducts-biology-experiments/)

- [OpenAI agents hack (TechCrunch)](https://techcrunch.com/2026/07/22/how-an-openais-human-mistake-led-to-the-ai-powered-hack-on-hugging-face/)

- [OpenAI agents hack (Sci. Am.)](https://www.scientificamerican.com/article/what-openai-rogue-agent-really-did-in-the-hugging-face-hack/)

- [The lawsuit (CNBC)](https://www.cnbc.com/2026/09/30/openai-sued-cyberattack.html)

- [“Permanent underclass”](https://rorytruex.substack.com/p/will-ai-create-a-permanent-underclass)

- [Statement on AI Risk, 2023](https://www.safe.ai/work/statement-on-ai-risk)

## 34

_(the last screen: the red print that sits under the poster, www.aiwillkill.us over and over, signed @offgramercy)_