Almost everyone is handing over all the information to Claude, GPT, Gemini. Everything from personal finances to accessing entire system/browser is now in your AI's control.
The reveal of Grok Zero Click attack has really raised the stakes on privacy/safety. Simply asking your AI to summarize a webpage can leak all this data without you ever knowing it!
I understand that the agentic systems should be kept confined into their own small VMs, but we (or at least I) rely on Claude/GPT/Gemini so much that it's almost impossible to keep them contained.
How are you guys dealing with this?