IP theft by AI models
Your input to AI can be used by it to train models and share the information with other users. This may not be a concern for most AI users who usually just ask questions and provide non-proprietary information. It can be a serious issue for technologists who use AIs to solve problems by providing proprietary context, such as source code.
To prevent IP theft by AI models:
ChatGPT:
Tap your profile name in the bottom left corner > Settings > Data Control > Improve the model for everyone > turn it off.
ChatGPT is good because it can use your context without using it to train its models.
Android Studio:
File > Settings > AI > Context Awareness > Use chat only, without project context
Gemini is worse. If it uses your context to solve a problem for you, it will use the it to train its models.