[scarmig]

I'm pretty sure I've already seen this in the wild, but there are plenty of workarounds.

1) Easiest is DOM-structure. Ads will often have a signature tell-tale placement and structure on a page, and this is enough to identify them. A naive ad implementation is trivial to detect, and even smarter ones can be worked around using crowdsourcing.

2) As a general matter, the facet of advertising we call ads are obvious to humans, and should in principle be detectable to machines. Rely on layout cues, other pages on the website, differences in content. This will ultimately push more and more advertising into the "sponsored content" category, which seems a genuinely harder problem.

[sangnoir]

The sponsored content/"native advertising" is the next logical step. It can range from "product placement" in otherwise regular articles, to full-blown advertorials disguised as regular articles. No DOM-structure to block, no layout clues and very non-obvious to humans. I consider these to be more insidious than current ads that most are eager to block

[dorgo]

What we see here is evolution of advertisement. From low quality spam-like ads to higher quality advertorials, which must provide some usefullness to the reader in order to carmouflage as normal content and to attract attention. I work in a marketing agency and there is a big shift from Adwords + SEO to "content marketing". Big money is put into creating interesting or funny stuff to attract attention (and backlinks).

[teddyh]

If the content is sponsored, this fact must still be disclosed according to FCC regulations. These disclosures could be detected. In fact, there is already a Firefox extension to do this: https://addons.mozilla.org/en-US/firefox/addon/addetector/