On the big platforms, the ads actually have “ad” written on them somewhere or “sponsored”.

It will be harder on other sites, but not that hard. There’s an unbreakable ceiling for the ads here - they need to be readable trivially by human eyes, or they don’t work as ads.

So, you can hide or obfuscate the ad… but then it’s not a very good ad.

Cost issues aside yeah image embedding tech is really good and can already do tasks like this, LLMs with native image understanding are even better at it.

Reasons this wouldn't work are you'd really need to get it to a good enough and small enough model to run low energy use and very fast to not impact browsing. The other issue is it'll be a fixed target for a period of time so the ads could be tweaked to find how to get them through it.

Cues, not queues. And yes that sounds workable.