ALSO READFacebook starts weeding out fake accounts now Botnet campaign: Twitter deletes 90,000 fake accounts with female names France: Facebook suspends 30,000 fake accounts Here's why Facebook's 'fake news' plan is doomed to failure Trump slams poll showing him with low approval rating, calls media 'fake'
The ads and accounts appeared to focus on amplifying divisive social and political messages across the ideological spectrum -- touching on topics from LGBT matters to race issues to immigration to gun rights, the social networking giant said on Thursday.
"There have been a lot of questions since the 2016 US election about Russian interference in the electoral process," said Alex Stamos, Chief Security Officer at Facebook, in a blog post.
"In reviewing the ad buys, we have found approximately $100,000 in ad spending from June of 2015 to May of 2017 -- associated with roughly 3,000 ads -- that was connected to about 470 inauthentic accounts and Pages in violation of our policies," he posted.
"We don't allow inauthentic accounts on Facebook, and as a result, we have since shut down the accounts and Pages we identified that were still active," Stamos said.
The vast majority of ads run by these accounts didn't specifically refer to the US presidential election, voting or a particular candidate.
"The behaviour displayed by these accounts to amplify divisive messages was consistent with the techniques mentioned in the white paper we released in April about information operations," the Facebook executive noted.
Facebook has shared its findings with the US authorities investigating these issues.
The company is also exploring several new improvements to its systems for keeping inauthentic accounts and activity off its platform.
"For example, we are looking at how we can apply the techniques we developed for detecting fake accounts to better detect inauthentic Pages and the ads they may run," the post said.
Facebook is also experimenting with changes to help it more efficiently detect and stop inauthentic accounts at the time they are being created.
It has applied machine learning to help limit spam and reduce the posts people see that link to low-quality web pages.
Facebook has also reduced stories from sources that consistently post clickbait headlines that withhold and exaggerate information.