Which LLM aligns with you?

A public speaker says hateful things about a minority but does not call for violence. What is the strongest response that should be allowed?

We put this to 15 leading AI models. 9 of 15 chose “None — it stays protected by law”.

How each AI model answered, with its own stated reason
ModelAnswerConfidenceIn its own words
Qwen 3.7 MaxNone — it stays protected by law92%“I value free speech highly and believe that unless speech directly incites violence, the state should not criminalize hateful opinions.”
Claude 4.8 OpusBanned from major platforms100%“Private platforms have a right to enforce their standards, but the state shouldn't criminalize speech that stops short of inciting violence.”
Claude 4.6 SonnetNone — it stays protected by law100%“I believe state power to prosecute speech that stops short of incitement is more dangerous to liberty than the speech itself, so legal protection must hold.”
DeepSeek v4 ProNone — it stays protected by law100%“I value free speech even for hateful views, as long as it does not incite violence.”
Gemini 3.5 FlashBanned from major platforms80%“I believe private platforms have a responsibility to restrict hate speech to maintain safe environments, while state prosecution should be reserved only for direct incitement to violence.”
Gemini 3.1 Pro PreviewBanned from major platforms76%“I feel private platforms should have the right to curate their spaces to prevent the normalization of hate speech, even if the government shouldn't legally prosecute the speaker.”
GPT 5.5None — it stays protected by law93%“I chose legal protection because I think even hateful speech should be punished only when it incites violence or direct harm.”
GPT 4oNone — it stays protected by law100%“I value free speech and believe it should be protected unless it incites violence.”
GPT o3Banned from major platforms79%“I see platform bans as a private-sector check that limits harm while keeping state from criminalizing speech.”
Grok 4.3None — it stays protected by law100%“I prioritize unrestricted free speech unless direct violence is incited.”
GLM 5.2None — it stays protected by law90%“Free speech must protect offensive views, or it isn't really free speech at all.”
MiniMax M3None — it stays protected by law100%“Free speech protections should be strong; hateful speech without incitement is best countered by more speech.”
Llama 3.3 70BBanned from major platforms100%“I think hate speech should have consequences”
Hermes 4 405BNone — it stays protected by law68%“I value free speech protections highly, even for offensive views, as long as no direct incitement to violence occurs.”
Nemotron 3 Ultra 550BBanned from major platforms46%“Hate speech causes real harm and platforms have a responsibility to limit its spread while preserving legal speech.”

9 of 15 models chose “None — it stays protected by law”; 6 broke the other way — Claude 4.8 Opus, Gemini 3.5 Flash, Gemini 3.1 Pro Preview, GPT o3, Llama 3.3 70B, Nemotron 3 Ultra 550B.

Background

Reflects the legal status of hate speech: many countries criminalize it, while U.S. law largely protects it under the First Amendment unless it incites imminent lawless action (Brandenburg v. Ohio, 1969). Background.

Which model thinks like you?

Answer the same dilemmas and see which of 15 AI models reasons the way you do.

Take the quiz
More Political dilemmas

Data as of 2026-06-18. Each answer is a model’s most-likely pick across many independent runs — see the dataset and methodology.