4 ms·

There are laws about disclosing ads, no machine learning required.

• username223 11y ago
  There is the CAN SPAM act, too, but it only works for legitimate organizations with reputations to preserve (politicians and legitimate retailers, mostly). SMTP-level blocking gets rid of most of the bottom-feeders, and ML targets the stuff in the middle. The web already has customer satisfaction and host-level blocking for the top and bottom, but ML will still be needed for the middle.
  • Nullabillity 11y ago
    The known sites aren't the one violating CANSPAM (and its ilk). There's a huge difference between websites that people choose to visit and an open inbox that anyone can ping.
    • username223 11y ago
      Given that even reputable/suable ad networks have been caught distributing malware, I don't see ad slots served by third parties as much better than open inboxes (though at least most mail clients don't load external resources or execute JS by default...). I think the web equivalent of CANSPAM-compliant mail is first-party advertising and third party networks who make a real effort to block malware and enforce some sort of design and privacy standards; maybe they would even obey do-not-track. The bottom-feeders are most of what currently shoves ads at people's faces.