Love this perspective! It’s wild to think our browsers are turning into secret agents. I’m curious, do you see any immediate tech solutions to stop AI from falling for these hidden prompt injecions?
Thank you. Great question — short answer: not yet.
Researchers are testing context firewalls (to block hidden prompts), source tagging (to verify where content comes from), and model-side filters for “instruction hygiene.”
But honestly, until these mature, the best safeguard is limiting what agentic AIs are allowed to automate.
I agree I think AI browsers are a bit premature for the majority of consumers.
Love this perspective! It’s wild to think our browsers are turning into secret agents. I’m curious, do you see any immediate tech solutions to stop AI from falling for these hidden prompt injecions?
Thank you. Great question — short answer: not yet.
Researchers are testing context firewalls (to block hidden prompts), source tagging (to verify where content comes from), and model-side filters for “instruction hygiene.”
But honestly, until these mature, the best safeguard is limiting what agentic AIs are allowed to automate.