Meta Unable to Block AI-Generated Hateful Political Ads during India's Election, New Investigation Finds

New Delhi: A new investigation by Eko, a corporate accountability organization has provided damming evidence of social media conglomerate Meta allowing hate-filled advertisements on its platform. Meta, which handles social media platforms such as Facebook and Instagram, reportedly approved 14 advertisements about India's election that had hateful content. 

Also Read | Smooth progress in testing latest HBM chips: Samsung

Eko, along with India Civil Watch International used AI to create advertisements that deliberately violated Meta's misinformation and hate speech policy. Of the 22 ads submitted to Meta's ad library, 14 were approved by Meta within 24 hours, the investigation found. Some of these ads contained inflammatory statements targeting certain communities and disinformation on political leaders. 

The investigation was conducted between May 8 to May 13. Despite Meta's claims that it is capable of filtering out content created through AI, Eko's investigation has revealed that the conglomerate did not block the provocative ads. A morphed video of Home Minister Amit Shah had gone viral on social media where he had allegedly threatened to cancel reservations given to the backward classes. Although some opposition leaders were severed notices and arrests were made, it was discovered that AI had been used to manipulate the video. 

Meta has been previously accused of inciting hate due to its lax content moderation in countries such as the USA and Brazil. The report also revealed that the 2020 Delhi riots, which killed 50 people saw Facebook encouraging hateful and violent content on its site.