The Natasha Doll controversy: how algorithms turned racism into a viral trend 📲

Cliquez ici pour lire en français

On Douyin and RedNote — China’s two dominant short-video platforms, with a combined user base of over 760 million — a new controversy is rattling social media. Videos out of China showing people physically abusing « Natasha Dolls, » a Black-skinned rubber toy, have triggered widespread outrage across African and Afro-descendant communities worldwide. For many users, these videos aren’t just a reflection of individual bad behavior: they normalize racism and help spread degrading stereotypes across the digital space. But beyond the outrage itself, what’s most striking is the speed at which recommendation algorithms pushed this content to millions of screens around the world.

A stress-relief toy at the heart of a disturbing trend ⚠️📸

The Natasha Doll is a soft, malleable rubber toy marketed as a stress-relief tool. It comes in multiple skin tones — Black, brown, and white — but it’s overwhelmingly the dark-skinned version that ends up at the center of the most violent videos. In widely shared clips, the dolls are punched, stomped on, run over, or doused with boiling water on camera. Propelled by sharing mechanics and algorithmic recommendation, these videos are reported to have racked up millions of views and fueled a trend that has sharply divided the internet.

Psychologists and content creators have also sounded the alarm about the trend’s potential to normalize aggression — particularly among younger viewers regularly exposed to this kind of content.

While some users have dismissed the clips as harmless entertainment, many others see them as a form of symbolic violence directed squarely at Black people. « This is mockery aimed at Black skin. It deserves to be sanctioned, » one commenter wrote.

The limits of content moderation 📱⚖️⚠️

This controversy has reignited a critical debate about the responsibility of digital platforms in allowing offensive content to circulate freely. Many observers are questioning whether current moderation systems — and the algorithms theoretically designed to detect and limit the spread of discriminatory content — are actually up to the task. The available data puts the scale of the challenge into sharp perspective: TikTok removed 204.5 million videos globally in the third quarter of 2025 alone — roughly 0.7% of all content uploaded — and acknowledges that 91% of those removals were carried out by automated systems. On X (formerly Twitter), the picture is even more revealing: of the more than 2 million hate speech posts removed in 2024, 99.75% required human moderators to intervene, with automated tools unable to independently handle the nuance and cultural context of discriminatory speech.

« Social media should primarily be about information and education. Today, some content seems to be doing the exact opposite — stoking controversy and driving confrontation, » one user observed.

This case unfolds against a broader global backdrop in which calls to combat online racial discrimination are growing louder. According to several observers, numerous human rights organizations and international institutions are pushing for stronger, more enforceable regulation of hateful content on digital platforms.

Accountability — who’s really responsible ? 🤖❓📱

For many users, the viral spread of these videos raises an uncomfortable question: has anyone actually been held accountable? « Since these Natasha Doll videos started circulating, has a single person been sanctioned? We’re talking about a country with a well-documented capacity for digital surveillance. The creators behind viral content could have been identified, » one user pointed out. The question feels even more pressing given that Douyin operates under direct oversight from the Chinese state — making the apparent absence of any enforcement action all the harder to explain.

Others prefer to put the incident in perspective, arguing that these videos reveal more about their creators’ ignorance than about any deeper social reality. Several commenters have reminded their audiences that « Africa exists and will continue to exist, regardless of these provocations. »

Taken together, the Natasha Doll affair illuminates one of the defining structural challenges of the digital age: the inability of algorithms and moderation systems to flag racially offensive content before it reaches scale. It also renews a broader and unresolved debate about the balance between freedom of expression, platform accountability, and the protection of communities targeted by harmful content — a debate that carries particular weight across Africa, where global platforms wield enormous influence over the information space without always investing moderation resources to match.

Do you think digital platforms are really doing their job when it comes to racist content? Let us know in the comments.