DIGITAL LIFE

Now under pressure, digital platforms create filters against ‘AI junk’

As “artificial intelligence junk” content spreads across the internet, efforts to contain the flood of images and videos considered low quality are also growing. Productions such as cats painting pictures, celebrities in compromising situations, or cartoon characters promoting products have become ubiquitous with the use of easily accessible AI tools, such as Google's Veo and OpenAI's Sora.

— The advancement of AI has generated questions about low-quality content, also known as AI junk [a term that has become popular in English as “AI slop”] — says Neal Mohan, CEO of YouTube.

According to critics, this is material created on a large scale, with little creative effort. This type of content is “cheap, bland and mass-produced,” says Swiss engineer Yves, who preferred not to give his last name, to AFP.

This assessment contrasts with the view of industry leaders. Satya Nadella, head of Microsoft, argues that the debate should be overcome and that technology should be adopted as an instrument to enhance creativity and productivity. The company is among the giants that invest heavily in artificial intelligence.

There are also those who see in the criticism of so-called "AI junk" a broader resistance to the democratization of creative tools.

— At its core, the criticism of AI junk is a criticism of individual creative expression — argues Bob Doyle, a YouTube influencer specializing in AI-generated content.

Stricter measures(below):

Despite the disagreements, digital platforms have begun to react. Pinterest informed AFP that it created a specific filter after receiving recurring requests from users who wanted to see fewer images of this type. TikTok introduced a similar feature at the end of last year.

YouTube, as well as Instagram and Facebook — both belonging to Meta — offer mechanisms to reduce exposure to this content, although they do not have explicit filters aimed solely at AI-generated productions.

Smaller companies are also adopting stricter measures. The music platform Coda Music, which has around 2,500 users, has begun allowing content created by artificial intelligence to be reported or even completely blocked.

"Until now, there has been a lot of participation in identifying AI artists," the company's CEO and founder, Randy Fusee, told AFP.

In the visual arts segment, the social network Cara, aimed at artists and designers and with over one million users, has implemented a combination of algorithms and human moderation to filter out AI-generated content. For its founder, Jingna Zhang, the users' demand is clear: "People want human connection."

The term "AI junk" (often called AI slop in English) refers to content (texts, images, videos) generated en masse by artificial intelligence that lacks effort, quality, or meaning, polluting the digital environment with clickbait and misinformation.

mundophone answers: Which platforms generate the most AI waste?...Based on recent reports (2025-2026), the main platforms where this content proliferates are:

YouTube: Reports indicate that more than 20% of the platform's content can already be considered "AI-generated junk," including fake videos, with synthetic avatars, or automatically generated to generate revenue.

TikTok: the platform generates a significant amount of "digital waste" and has a considerable environmental impact, mainly due to the high energy consumption required to maintain its data centers and the data traffic of short videos...Currently, it is adopting some measures to try to reduce digital waste.

Facebook and Meta (Threads/Instagram): There has been a massive increase in AI-generated posts, with researchers pointing out that the algorithm itself sometimes boosts this content, which includes fake images of political figures and curiosity clicks.

Pinterest: Users report that the visual inspiration platform has been "flooded" with AI-generated images, leading the platform to create tools to "tune" (adjust/reduce) the amount of AI in feeds.

Music Platforms (e.g., Deezer): Streaming services are facing a flood of AI-generated music (tools like Suno and Udio), with reports of over 60,000 automatically uploaded tracks daily, often by fake accounts.

Amazon Kindle and E-books: A notable increase in AI-generated digital books with fictitious authors and low quality, forcing platforms to create publishing limits.

LinkedIn: The network has seen an increase in generic, low-quality posts, automatically produced by content automation tools that scrape websites and generate posts with little "human touch." Main Tools/Sources of AI Slop:

Text and image generators: Copy.ai, Jasper.ai, Writesonic, Midjourney, DALL-E, Canva.

Video/Avatar Tools: Synthesia, InVideo.

Audio/Music Tools: Suno, Udio.

Context of the "War on AI Slop"

In 2026, platforms like YouTube intensified the fight against this material, trying to balance the use of AI for creators with combating low-quality automated content. Meanwhile, new platforms focused on human content (such as Cara and Pixelfed) are emerging as "anti-AI" alternatives.

mundophone