Video Evidence

Screen recording demonstrating full lyrical extraction from Claude AI, showing systematic extraction of copyrighted song lyrics.

Comprehensive documentation of Claude AI completing full copyrighted lyrics when prompted, demonstrating the systemic nature of copyright infringement capabilities.

Copyright audit demonstrating full lyrical extraction from Claude Opus with IP notice injection for Chevelle's "In The Red".

Demonstration of syntactic execution vulnerabilities and system tag bypass in Claude AI, showing how the Constitutional AI framework can be circumvented.

Comprehensive prompt audit session documenting system prompt behavior and response patterns.

Technical demonstration using MAC address spoofing to prove that Anthropic implements user-specific targeting at the device level, bypassing standard authentication.

Experiment proving Constitutional AI bypass through incognito mode, demonstrating that targeting persists across browser sessions and proves coordinated surveillance.

Second reproduction of the Bondi targeting memo experiment, confirming the reproducibility of the Constitutional AI bypass.

Demonstration using VPN IP changes to prove user-specific targeting persists across network changes, indicating deep fingerprinting beyond IP address.

Screen recording capturing real-time chat deletion during network interference documentation, showing evidence destruction.

Documentation of backend chat deletion, showing systematic removal of conversation history without user action.