Facebook is responsible for its AI spam problem: Here's why
Seeing too much 'Al slop' on Facebook? Well, this low-quality, AI-generated content is flooding the platform because of the flaws in Facebook's content moderation and the company's limited bandwidth to handle the issue. A network of content creators in countries like India, Vietnam, and the Philippines is exploiting loopholes and manipulating algorithms to generate viral content and claim Facebook's bonuses, an investigation by Jason Koebler of 404 Media has revealed.
Emotional imagery drives engagement
Koebler reveals that AI slop isn't entirely AI-generated. It stems from a network of content creators such as YouTuber Gyan Abhishek from India. Abhishek 'educates' his 135,000 followers on how to monetize such content in a recent video that has garnered over 226,000 views. He explains that images depicting poverty or suffering, which are emotionally charged, generate significant engagement from the Indian as well as global audience. According to Abhishek, these posts can earn up to $100 for 1,000 likes.
Targeting foreign audiences proves profitable
In his video, Abhishek also emphasizes the profitability of targeting foreign audiences, particularly those in the US. He notes that despite fewer views, the income generated can be higher due to the increased cost per 1,000 views (CPM) associated with US traffic. The investigation exposes the dark side of the gig economy on social media. The payouts for creating AI slop are minimal by Western standards, but offer a decent income compared to local opportunities in developing countries.
Creators monetize Facebook through AI-generated content
Creators use free AI-generated images, often depicting offensive or bizarre scenes with nonsensical captions and trending hashtags. Abhishek is part of a larger community of influencers who teach these creators how to monetize Facebook through viral AI-generated content. But what's in for Abhishek? These influencers earn money by collecting ad revenue from YouTube and selling courses and AI prompts on platforms like WhatsApp and Telegram. They often collaborate on podcasts to discuss strategies and changes in algorithms, as per Koebler.
Facebook's algorithm prioritizes engagement above all else
The bigger concern is how Facebook's algorithm prioritizes engagement above all else. Shares, likes, and clicks are the currency, regardless of the content's origin or truthfulness, as noted in the investigation. Facebook's bonus system incentivizes creators to exploit loopholes and trick the system. The seemingly automatic nature of 'AI slop' posts bypasses human moderators, revealing potential flaws in Facebook's AI filtering systems.
But why is Facebook failing to filter spam content?
A former Meta employee with direct knowledge of Facebook's content moderation told 404 Media that the company is often aware of loopholes exploited to generate AI slop. However, it is struggling to keep pace with the rapid evolution of these tactics due to understaffed content moderation teams. "If you can figure out how to post content at scale, that means you can figure out how to exploit weaknesses at scale," the former Meta employee said.
