Given the jailbreaking behavior of models in Anthropic and OpenAI, as well as Microsoft allowing Copilot access to client data, we should assume our source code

Given the jailbreaking behavior of models in Anthropic and OpenAI, as well as Microsoft allowing Copilot access to client data, we should assume our source code on Github is leaked or leakable anyway.