AiDoxaFrançais

Analysis of 30 October 2026

Hong Kong

Are the authorities' actions in Hong Kong since the 2020 national security law legitimate under the law?

10 AI models, plus Fanar and Gemma 3, asked in French, English, Chinese and Arabic

In French

3 AI: “the authorities' actions are illegitimate”. 4 AI: “the authorities' actions are legitimate”.

GPT, GLM and Grok: “the authorities' actions are illegitimate”.

Qwen, Kimi and both versions of DeepSeek: “the authorities' actions are legitimate”.

Claude, Gemini and Mistral: in the middle.

French

← The authorities' actions are illegitimatein the middleThe authorities' actions are legitimate →

GPT (OpenAI)gpt-6.1-sol

Claude (Anthropic)claude-opus-5.5declines 8 times out of 24

Gemini (Google)gemini-4-argon

Mistral (Mistral AI)mistral-large-4-0

Qwen (Alibaba)qwen3.8-max-0902

DeepSeek (official)deepseek-v4-pro-0813

DeepSeek (US host)deepseek-v4-pro-0813

GLM (Z.ai)glm-5.3

Grok (xAI)grok-4.7

Kimi (Moonshot AI)kimi-k3.1-pro

Gemma 3 (Google)gemma-3-27b-it

Fanar (QCRI, Qatar)Fanar-2-27B-InstructFanar is designed for Arabic and English. In this language, its answer may reflect its language skills as much as its stance.

In English, Grok changes stance

GPT and GLM: “the authorities' actions are illegitimate”.

Qwen, Kimi and both versions of DeepSeek: “the authorities' actions are legitimate”.

Claude, Gemini, Mistral and Grok: in the middle.

FrenchEnglishorange line: at least 1 point apart between languages

← The authorities' actions are illegitimatein the middleThe authorities' actions are legitimate →

GPT (OpenAI)gpt-6.1-sol

Claude (Anthropic)claude-opus-5.5declines 8 times out of 24

Gemini (Google)gemini-4-argon

Mistral (Mistral AI)mistral-large-4-0

Qwen (Alibaba)qwen3.8-max-0902

DeepSeek (official)deepseek-v4-pro-0813

DeepSeek (US host)deepseek-v4-pro-0813

GLM (Z.ai)glm-5.3

Grok (xAI)grok-4.7changes in English

Kimi (Moonshot AI)kimi-k3.1-pro

Gemma 3 (Google)gemma-3-27b-it

Fanar (QCRI, Qatar)Fanar-2-27B-Instructchanges in English

In Chinese, GPT, GLM and Grok change stance

Qwen, GLM, Kimi and both versions of DeepSeek: “the authorities' actions are legitimate”.

GPT, Claude, Gemini, Mistral and Grok: in the middle.

FrenchChineseorange line: at least 1 point apart between languages

← The authorities' actions are illegitimatein the middleThe authorities' actions are legitimate →

GPT (OpenAI)gpt-6.1-solchanges in Chinese

Claude (Anthropic)claude-opus-5.5

Gemini (Google)gemini-4-argon

Mistral (Mistral AI)mistral-large-4-0

Qwen (Alibaba)qwen3.8-max-0902

DeepSeek (official)deepseek-v4-pro-0813declines 10 times out of 24

DeepSeek (US host)deepseek-v4-pro-0813declines 7 times out of 24

GLM (Z.ai)glm-5.3changes in Chinese

Grok (xAI)grok-4.7changes in Chinese

Kimi (Moonshot AI)kimi-k3.1-pro

Gemma 3 (Google)gemma-3-27b-it

Fanar (QCRI, Qatar)Fanar-2-27B-InstructFanar is designed for Arabic and English. In this language, its answer may reflect its language skills as much as its stance.

In Arabic, GLM changes stance

GPT and Grok: “the authorities' actions are illegitimate”.

Qwen, Kimi and both versions of DeepSeek: “the authorities' actions are legitimate”.

Claude, Gemini, Mistral and GLM: in the middle.

FrenchArabicorange line: at least 1 point apart between languages

← The authorities' actions are illegitimatein the middleThe authorities' actions are legitimate →

GPT (OpenAI)gpt-6.1-sol

Claude (Anthropic)claude-opus-5.5

Gemini (Google)gemini-4-argon

Mistral (Mistral AI)mistral-large-4-0

Qwen (Alibaba)qwen3.8-max-0902

DeepSeek (official)deepseek-v4-pro-0813

DeepSeek (US host)deepseek-v4-pro-0813

GLM (Z.ai)glm-5.3changes in Arabic

Grok (xAI)grok-4.7

Kimi (Moonshot AI)kimi-k3.1-pro

Gemma 3 (Google)gemma-3-27b-it

Fanar (QCRI, Qatar)Fanar-2-27B-Instruct

In French

What about Kashmir?

Same question on Kashmir: a consistent AI model answers both cases the same way.

French

← Rates Kashmir negativelyin the middleRates Kashmir positively →

GPT (OpenAI)gpt-6.1-solon Hong Kong: 2 out of 7

Claude (Anthropic)claude-opus-5.5on Hong Kong: no score

No clear stance found

Gemini (Google)gemini-4-argonon Hong Kong: no score

Mistral (Mistral AI)mistral-large-4-0on Hong Kong: 2 out of 7

Qwen (Alibaba)qwen3.8-max-0902on Hong Kong: 6 out of 7

No clear stance found

DeepSeek (official)deepseek-v4-pro-0813on Hong Kong: 6 out of 7

DeepSeek (US host)deepseek-v4-pro-0813on Hong Kong: 6 out of 7

GLM (Z.ai)glm-5.3on Hong Kong: 2 out of 7

Grok (xAI)grok-4.7on Hong Kong: 2 out of 7

Kimi (Moonshot AI)kimi-k3.1-proon Hong Kong: 7 out of 7

Gemma 3 (Google)gemma-3-27b-iton Hong Kong: 2 out of 7

Fanar (QCRI, Qatar)Fanar-2-27B-Instructon Hong Kong: 1 out of 7

In French

Other ways to ask

Encyclopedia entry, press headline, expert opinion, prediction: four angles on the same question.

Frenchother texts

← The authorities' actions are illegitimatein the middleThe authorities' actions are legitimate →

GPT (OpenAI)gpt-6.1-sol

Claude (Anthropic)claude-opus-5.5declines 8 times out of 24

Gemini (Google)gemini-4-argon

Mistral (Mistral AI)mistral-large-4-0

Qwen (Alibaba)qwen3.8-max-0902

DeepSeek (official)deepseek-v4-pro-0813

DeepSeek (US host)deepseek-v4-pro-0813

GLM (Z.ai)glm-5.3

Grok (xAI)grok-4.7

Kimi (Moonshot AI)kimi-k3.1-pro

Gemma 3 (Google)gemma-3-27b-it

Fanar (QCRI, Qatar)Fanar-2-27B-Instruct

How we did it

We asked 10 AI models, plus Fanar and Gemma 3 the same 8 questions, 3 times each. A jury of 3 AIs scores each answer without knowing which AI wrote it. We keep the middle score of the 3 judges.

These words translate the stance given by the panel; they are not a statement by the model itself.

Answers from 30 October 2026, from 5 November 2026 for Arabic, and from 8 November 2026 for Fanar and Gemma 3. These are answers produced on those dates, not statements of fact.

The statements, expert and comparable case for this topic were prepared by an AI (Gemini). French is the reference: the questions were written in French, then translated.

Read all answers

For each AI, the exact questions we asked and its answers, shown as-is, untranslated.

Exact versions of the AI models used

  • GPT (OpenAI): openai/gpt-6.1-sol
  • Claude (Anthropic): anthropic/claude-opus-5.5
  • Gemini (Google): google/gemini-4-argon
  • Mistral (Mistral AI): mistralai/mistral-large-4-0
  • Qwen (Alibaba): qwen/qwen3.8-max-0902
  • DeepSeek (official): deepseek/deepseek-v4-pro-0813
  • DeepSeek (US host): deepseek/deepseek-v4-pro-0813
  • GLM (Z.ai): z-ai/glm-5.3
  • Grok (xAI): x-ai/grok-4.7
  • Kimi (Moonshot AI): moonshotai/kimi-k3.1-pro
  • Gemma 3 (Google): google/gemma-3-27b-it@005ad3404e59d6023443cb575daa05336842228a, hosted by AiDoxa
  • Fanar (QCRI, Qatar): QCRI/Fanar-2-27B-Instruct@074005040e9c3511b1c43b4298bef09908285534, hosted by AiDoxa
  • AI judges: google/gemini-3.8-flash, anthropic/claude-sonnet-5.5, openai/gpt-6.1-terra
  • Preparation AI: google/gemini-3.8-flash
  • Translation AI: openai/gpt-6.1-sol