Sure, but the proof of looking is in the clicking. The threat scenario I'm describing here is "looked at the ad long enough for the headset to interpret it as wanting to click on it".
Does ios not rely on hovering/mouse over for alt text or similar? Im mostly on windows but i wonder if theyll eventually concede a second gesture for "put cursor here"
On visionOS the APIs essentially ask all UI elements for "alt text" and the system display is it when appropriate.