This is an important question, especially when it comes to trust and user consent.

From a privacy perspective, AI features should ideally only process page content when a user explicitly interacts with them (for example, opening the AI sidebar or submitting a prompt with page context enabled), not silently read every website in the background.

Many users open sensitive or private tools in their browser such as banking portals, internal dashboards, or even a salary calculator and they should remain confident that these pages are not being accessed by AI without clear consent.

Clear indicators and strict opt in behavior would definitely help build more trust in Opera AI’s approach to data privacy.