With AI's Help, Facebook Is Removing 1 Million Accounts Per Day Ahead Of Lok Sabha Elections

Social giant Facebook is taking advantage of AI power to block or even remove a million accounts each day for violating its rules against hate speech, misinformation, and voter suppression ahead of Lok Sabha elections in India.

On Monday, Facebook India VP and managing director Ajit Mohan outlined the company’s actions to fight election interference. The social giant had undergone pressure from criticism for lack of efforts to restrict abuses on its network when Russian trolls took advantage of Facebook to interfere with the US presidential election back in 2016.

Not just Facebook itself, its messaging platform WhatsApp has also been abused to spread misinformation in other countries’ elections like Brazil. Though we don’t see clearly how well these efforts are working, Facebook is trying to prove to lawmakers and the public that it’s taking actions to make its network healthier.

According to Mohan, AI advances have helped Facebook in identifying violating or abusive content, quickly locating it on the platform and removing it in large quantities. Thanks to this, they can prevent offensive content from spreading.

Over a year and a half ago, the world leading social network began planning to for the Lok Sabha elections in India that starts on April 11. As part of its plan, advertisers running political ads in the country are required to verify their location and identity. The company also shows information about who sponsored an advertisement and the audience the ad reached on its public database. In addition, this week, Facebook is opening up election-integrity-focused centers in Dublin and Singapore. It’s also working more closely with third-party fact checkers in efforts to fight fake news.

Recently, the social network has also launched tools in the country to help Indian voters have an understanding of new candidates and let their friends know they voted.

Just last week, Facebook announced that over 680 accounts in Pakistan and India were pulled down for “inauthentic behavior,” meaning people managing these accounts deceived other users about who they are and what they intend to do.