I'm not aware of any similar arguments being tested in court, so this is all conjecture.

That said, barring an explicit definition of 'Internet service' in your service contract, it's commonly understood that requesting a page from example.com, all the data your ISP returns implicitly is sourced from example.com. Introducing your own content in between is therefore fraud, as you've mis-represented the origin of the content.

I believe the owner of an involved web server would have standing as well, not just the users.

But in such a case you could make the same argument for ad supported wifi connections or proxies.

You do enter into agreement (usually), but do you have a valid contract when using those connections?

Maybe we should all start packing our own TOS's in request headers... :)

FWIW, I haven't noticed this on any of BT, Sky or Virgin.

NXDOMAIN is hijacked on Virgin Media for sure, unless you are using some other DNS. I haven't noticed them injecting ads though.

Not to be needlessly contrarian, but this doesn't appear to be the case - at least for me, for any of the domains I tried:

>nslookup notarealdomainthatshouldresolve.co.uk Server: cache1.service.virginmedia.net Address: 194.168.4.100

* cache1.service.virginmedia.net can't find notarealdomainthatshouldresolve.co.uk: Non-existent domain

NXDOMAIN used to be hijacked by Virgin Media. But it certainly is not any more, unless their DNS service changes significantly across the country.