The number of ‘bad ads’ taken down by Google fell to 2.3bn in 2018, almost 1bn less than it removed the year prior. In its latest ‘Trust & Safety Ads Report’, the internet giant said that these ads included almost 207,000 ads from ticket resellers, more than 531,000 for bail bonds, and around 58.8m phishing ads.

Although Google purged fewer ads in 2018, it increased the number of publishers and advertisers that it terminated. The search engine booted nearly 1m advertiser accounts, which is close to double the amount terminated in 2017, while increasing the number of publishers and app developers it removed from 320,000 to 734,000. Moreover, it completely removed ads from almost 1.5m apps and did the same for nearly 28m pages for violating its publisher policies.

To help identify the pages violating its policies, Google launched 330 detection classifiers in 2018 – almost three times the number it introduced in 2017 – to help the company better detect ‘badness’ at the page level.

When it came to addressing what is arguably the biggest challenge facing Google and other major tech companies, political advertising, the company launched policies for election ads ahead of the US midterm elections. Through these policies, Google verified around 143,000 election ads and introduced a report to provide information about those buying election ads. This year, Google will be launching similar tools ahead of key elections in the EU and India.

Google also took down around 1.2m pages, over 22,000 apps, and close to 15,000 sites for spreading “misrepresentative, hateful or other low-quality content”.