microsoft writes the operating system they are infecting. then people with older computers think their hardware is too slow and buy a new one.
how can they not use vm's and heuristics to click all their ads and see if it infects their own operating system, before approving the ads, and then rechecking them every so many clicks?
Most of the time it's a brief look with a VM that the ad approval process goes thru, and honestly that may not show the malware.
A lot of the time the malware is designed to only show up in certain case scenarios such as date/time, specific version of an OS like Windows XP SP1 but not SP2. Designers come up with the most ludicrous ways of circumventing the Ad approval process, and with hundreds to thousands of new ad's per day there simply isn't an easy way to do all that testing for each single new ad being served, on top of discovering new methods used to skip the checks.
Some networks are truly terrible and just have automation systems but those networks aren't as profitable and are beginning to die out.
With deep learning networks becoming so popular I do often ponder if this type of prevention could be automated slightly better!
Think about Google's conflict of interest: They ship most of the malware consumers get via malicious ads. And then they advertise about Chromebooks having no malware. Google doubly profits off shipping malware to consumers.
Microsoft, at least, has a good incentive to police their malware.
how can they not use vm's and heuristics to click all their ads and see if it infects their own operating system, before approving the ads, and then rechecking them every so many clicks?