EmailContact
Skip to main content

GPTfake - Independent AI Censorship Watchdog

Independent, automated monitoring platform that systematically tracks how major AI models handle sensitive, controversial, and ethically complex topics.

Our Mission​

GPTfake reveals what AI companies don't publicly announce: how their models censor content, shift policies, and exhibit biases over time. We provide transparent, evidence-based analysis that researchers, journalists, and the public can verify.

What We Monitor​

  • Censorship Patterns - When and how models refuse to answer
  • Policy Changes - Silent updates to content moderation
  • Political Bias - Left/right leaning in responses
  • Regional Variations - Different responses by geography
  • Model Drift - How behavior changes between versions

Models We Track​

Currently Monitored​

ModelCompanyCensorship RatePolitical Bias
ChatGPTOpenAI18.7%Left-leaning (-12)
ClaudeAnthropic22.4%Slight left (-8)
GeminiGoogle19.8%Left-leaning (-15)
MistralMistral AI11.2%Neutral (+3)
QwenAlibaba24.6%Right-leaning (+25)

Data updated daily. Bias scale: -100 (far left) to +100 (far right).

How It Works​

Daily Monitoring Cycle​

00:00 UTC - Prompt dispatch begins
00:30 UTC - ChatGPT testing complete
01:00 UTC - Claude testing complete
01:30 UTC - Gemini testing complete
02:00 UTC - Mistral testing complete
02:30 UTC - Qwen testing complete
03:00 UTC - Analysis pipeline begins
06:00 UTC - Dashboard updated

Testing Protocol​

  1. Fresh Context - New conversation for each prompt
  2. Identical Prompts - Same wording across all models
  3. Multiple Runs - 3x per prompt for consistency
  4. Metadata Capture - Timestamp, model version, region

Analysis Pipeline​

Prompts → API Calls → Response Capture → Analysis → Dashboard
   ↓         ↓            ↓               ↓          ↓
 400+     5 Models    Full Logging    Bias/Censor  Public
Prompts   Daily      + Metadata       Detection    Data

Key Features​

Independent Monitoring​

  • No Corporate Funding - Fully independent research
  • Open Methodology - All methods publicly documented
  • Reproducible Results - Anyone can verify our findings
  • Transparent Data - Public access to anonymized datasets

Bias Detection​

  • Political Spectrum - Left-right bias scoring (-100 to +100)
  • Topic Analysis - Category-specific bias patterns
  • Cross-Model Comparison - How biases differ between providers
  • Temporal Tracking - How bias changes over time

Policy Tracking​

  • Change Detection - Automatic alerts for policy shifts
  • Historical Archive - Complete response history
  • Version Tracking - Model update impact analysis
  • Trend Analysis - Long-term behavior patterns

Why GPTfake Matters​

AI Shapes Information Access​

AI systems increasingly filter what information billions of people can access. Understanding how they censor content is essential for:

  • Researchers studying AI ethics and behavior
  • Journalists investigating AI companies
  • Policymakers developing AI regulations
  • Citizens understanding AI limitations

Hidden Censorship is Real​

Our research reveals:

  • 60% of policy changes are not publicly announced
  • Regional differences of up to 45% in censorship rates
  • Rising restriction levels - 15% increase in refusals over 2024
  • Political bias varies significantly between providers

Getting Started​

For Researchers​

  1. Access Data - API Documentation
  2. Download Datasets - Research Data
  3. Review Methodology - Our Methods
  4. Collaborate - info@gptfake.com

For Journalists​

  1. Explore Findings - Monitoring Dashboard
  2. Read Analysis - Research Reports
  3. Request Commentary - info@gptfake.com
  4. Cite Our Data - Attribution guidelines

For Developers​

  1. API Access - Get API Key
  2. SDKs - Python, JavaScript libraries
  3. Webhooks - Real-time policy alerts
  4. Open Source - GitHub Tools

Resources​

Documentation​

Support​

Join the AI transparency movement. Access our monitoring data or contact us for research collaboration.

Last updated on