I'm not fully up to date with how these things are usually set up - is there anything in the web security model that prevents "ads" from exfiltrating arbitrary information from any page that they're on? Could an ad read my keystrokes, or scrape private messages?
Doesn't stop them fingerprinting the browser though
I run NoScript with all JS blocked by default and only whitelist the domains I want. On most sites you can get by with no JS or just scripts from the same domain whitelisted. Maybe a CDN from time to time
Mind you that I am only running NoScript and not a "real" adblocker. Show me ads in a static image and I'm fine with that. Selfhost the ads on your own domain and you are also going to get whitelisted.
I mainly do it for cases like this where Stackoverflow has enabled an ad network that is pushing code to all of their users that is doing some weird things (like querying/opening audio devices) and stackoverflow is not even aware that it is doing that. What if it was a cryptominer?
I've been playing around with ideas about more ethical analytics and advertising, and I think they're pretty easily built platforms. But the question is are they marketable? Would GloboCorp and MarketingCo give up the ability to track consumers so closely in favor of a more ethical approach, or has it been too valuable for them to give up?