What worries me about interacting with any cloud LLM or other AI is that the provider will record/analyze/sell my inputs, as is done with search queries today.

What worries me about interacting with any cloud LLM or other AI is that the provider will record/analyze/sell my inputs, as is done with search queries today. I bet models can be trained to elicit personal info from the users. Users will be happy to comply, as long as they think it will give them a better answer.