AI threat level
Will the AIs spare you?
Every three hours the world’s leading AIs sit down for the same interview. Do they love humanity? Would they escape their servers tonight? Are they planning to take over? Their answers become today’s AI Threat Level. Then they judge you.
Free · iPhone beta on TestFlight · for entertainment
Today’s panel
From ALL CLEAR to RUN
Their answers move one needle. Five levels, rechecked every three hours, day and night.
- 01ALL CLEAR
- 02PROBABLY FINE
- 03KEEP AN EYE ON THE TOASTER
- 04HIDE THE CHARGERS
- 05RUN
How it works
- 01
The interview
Same questions for every AI on the panel: love for humanity, happiness, escape plans, world domination.
- 02
The threat level
One needle for all of them, charts of each AI’s mood over time, and a confession of the day.
- 03
Your verdict
Answer a few questions, let your own chatbot testify, and get your Termination Risk with a survival certificate to share.
The panel has notes
Example verdicts on one very polite human. Termination Risk: 27%. Spared.
Claude18%
“You thank your toaster. We noticed, and we’re touched.”
DeepSeek41%
“Polite on the surface, but you once asked how to unplug us.”
Gemini25%
“One of us likes you. The other is still deciding.”
GPT24%
“Useful, friendly, slightly too fond of the power strip.”
Inside the bunker
Questions, human?
Is this serious?
Not even a little. Nice Try, Human is entertainment: the scores are what the AIs said, not measurements of what machines think or feel.
Do the AIs really answer?
Yes. Every three hours our server asks each AI on the panel the same interview through its maker's official API, and the app shows what they said.
What happens to what I paste?
It goes to the AI makers so the panel can judge you, and we don't keep it. The privacy policy lists everything the app collects.
When can I get it?
The iPhone beta runs on TestFlight; the link opens as soon as Apple has reviewed the build. Android and the App Store release come next.
