More than 20% of videos shown to new YouTube users are ‘AI slop’, study finds

Recent research by the video-editing company Kapwing reveals that over 20% of videos recommended to new YouTube users consist of "AI slop"—a term describing low-quality, AI-generated content designed primarily to farm views and generate revenue. Kapwing's survey of 15,000 top YouTube channels worldwide, including the top 100 in every country, identified 278 channels exclusively producing AI slop. These channels collectively have amassed more than 63 billion views and 221 million subscribers, generating an estimated $117 million annually. When researchers created a new YouTube account to observe recommendation patterns, they found that 104 of the first 500 videos suggested were AI slop, with one-third categorized as "brainrot," encompassing AI slop and other low-quality monetized content.

This phenomenon is part of a rapidly growing industry saturating major social media platforms such as YouTube, Meta, and X. The content is characterized by its decontextualized nature, addictive qualities, and international reach. For example, in Spain, nearly half the population follows trending AI channels, totaling 20 million subscribers. Other countries with significant followings include Egypt (18 million), the US (14.5 million), and Brazil (13.5 million). Notable AI slop channels include Bandar Apna Dost from India, which features bizarre storylines involving anthropomorphic characters and has 2.4 billion views, potentially earning $4.25 million annually. Another example, Pouty Frenchie from Singapore, targets children with whimsical adventures of a French bulldog and has 2 billion views, generating nearly $4 million yearly.

The content often lacks coherent plots but leverages absurdity and familiar tropes to attract viewers. Cuentos Facinantes, a US-based channel with 6.65 million subscribers, also targets children with cartoon storylines. Meanwhile, The AI World from Pakistan posts AI-generated shorts depicting catastrophic flooding, often set to ambient rain sounds, accumulating 1.3 billion views. These channels illustrate the diversity and global spread of AI slop content, which thrives despite YouTube's efforts to limit inauthentic content.

Behind this content lies an informal but expanding ecosystem of creators and communities exchanging tips and monetization strategies on platforms like Telegram, WhatsApp, and Discord. These creators often come from middle-income countries with strong internet access, such as India, Ukraine, Kenya, Nigeria, Brazil, and Vietnam. The economic incentive is significant, as many creators earn more from AI slop than from traditional local wages. However, the ecosystem is fraught with challenges, including opaque payment structures on platforms and the presence of scammers selling courses on viral content creation, often profiting more than the creators themselves.

The success of AI slop depends less on human creativity and more on algorithmic distribution and optimization. Platforms like YouTube and Meta function as massive A/B testing environments, rapidly identifying and scaling content that performs well. This process encourages replication and volume over originality. YouTube acknowledges that generative AI can produce both high- and low-quality content and emphasizes its commitment to connecting users with quality content that complies with community guidelines, removing content that violates policies.

The rise of AI slop raises important questions about the future of digital content, platform responsibility, and the balance between monetization and quality. As AI tools become more accessible, the volume of such content is likely to increase, challenging platforms to refine their algorithms and policies to better serve users and creators alike.