The company hasn't outlined the full scope of its ad removals, but software threads only played a small part. Out of the 3.2 billion removed ads, 'only' 79 million were yanked for steering people to malware sites. Google also pulled 66 million "trick to click" ads (that is, they masquerade as something besides an ad) and 48 million ads goading people into installing unwanted software. To put it another way, ad removals weren't focused on active threats to your devices.
It's tempting to suggest that fake news and misinformation campaigns played a large part, but it's not clear that this is the case. Google, at least, downplays it -- the company claimed there were a "small number" of ad publishers to blame for "misrepresentative content" like fake news or scams. Out of 11,000 potential policy violators, it only found 90 publishers and 650 sites that warranted a crackdown. Instead, Google saw a larger problem with sites that copy content (such as, say, news articles). It blocked more than 12,000 of those sites.
Still, it's evident that Google was much more aggressive in tackling bad ads throughout 2017, in no small part due to government pressure. Politicians and regulators want to know that Google isn't aiding Russian meddling, fake news purveyors and scams, and aggressive clean-up efforts like this could show that Google is serious enough about fighting bad ads to avoid government intervention.