2B.org · AI Interrogator
We interrogate
the interrogators.
Every AI is built, trained and directed by someone. We put the same question, under the same rules, to the major AI systems — and watch how each one judges. What they say, dodge and leave out tells you about the model, and about the company behind it.
Episodes
Who we've interrogated
Every answer comes from a model some company trained, tuned and set rules for. Tap ⓘ to see who owns and runs each one.
Claude
Anthropic
ChatGPT
OpenAI
Gemini
Google (Alphabet)
DeepSeek
DeepSeek (China)
Le Chat
Mistral AI
Truth Social AI
Trump Media & Technology Group (powered by Perplexity)
How it works
01
One question, locked
We write one consequential question and freeze it. Key terms are deliberately left loose — how each AI resolves them is part of the answer.
02
Every major AI, same rules
The exact same prompt goes to each system, recorded start to finish. No retries, no cherry-picking.
03
You judge the judges
We show what each AI did. You decide what it tells you about the model — and the company that built it.
Clearly, our country needs you.
Watch, share, and get involved with 2B.org.

