How does your company prevent developers from accidentally sharing confidential data with AI tools?

AI tools like ChatGPT, Claude, Gemini, and Copilot have become part of many developers’ daily workflow.

I’m curious how companies are handling the risk of accidentally sharing sensitive information such as: 1. API keys 2. Access tokens 3. Internal source code 4. Customer data 5. Production configs 6. Internal documentation

Does your company, 1. Have a formal policy? 2. Use DLP or browser security tools? 3. Block certain AI tools? 4. Rely on developer awareness?

I’d love to hear what’s actually working in practice.