AI-Generated Black Female Avatars Spark Racism Concerns Across Social Media

TikTok has removed 20 accounts following a BBC inquiry that revealed a troubling trend of AI-generated black female avatars deployed to drive users towards sexually explicit content. The platform took action after the BBC and experts at the AI research outlet Riddance discovered dozens of accounts across Instagram and TikTok featuring heavily sexualised digital characters with exaggerated body shapes and darkened skin. The accounts, which were not labelled as artificially created in apparent breach of platform policies, used racist tropes in their naming practices and marketing copy. Whilst TikTok has acted swiftly, Meta, Instagram’s parent company, said it was looking into but has not confirmed taking comparable steps against the accounts active on its platform.

The Finding: Many Misleading Accounts

The BBC’s investigation, carried out alongside Riddance researchers Jeremy Carrasco and Angel Nulani, uncovered a sophisticated operation spanning multiple platforms. The team identified 60 accounts mainly found on Instagram that held links that directing users to commercial adult content on external sites. Notably, whilst these third-party sites described the images as AI-generated, the Instagram accounts themselves made no such disclosure, creating a deliberate deception for unsuspecting users. The research exposed an substantially wider collection of similar accounts on Instagram and TikTok alike that contained no links to commercial material, indicating the problem extends far beyond commercial exploitation.

The accounts employed a deliberate approach to evade detection and expand reach quickly. The vast majority of accounts were focused on Instagram, with approximately one-third also operating accounts on TikTok. Account names intentionally employed racially charged language, including terms such as “black”, “noir”, “dark” and “ebony”, alongside language promoting racial prejudices and objectification. Numerous accounts followed or engaged with one another, forming a coordinated web that amplified their reach and exposure. This organised strategy points to intentional organisation rather than isolated incidents, highlighting critical issues about the scale and sophistication of the operation.

• AI avatars featured distorted physical proportions and darkened skin tones
• Account names contained racial terminology and stereotypical language about white men
• Videos were not labelled as AI-created, contravening platform guidelines
• Many accounts engaged in mutual following, forming coordinated systems for reach enhancement

Abuse Via Synthetic Generation

Unauthorised Material and Online Alteration

The investigation revealed a particularly troubling facet to the enterprise: the systematic appropriation of creative work from genuine creators. One profile that gathered three million followers within weeks of its December launch had methodically stolen footage from genuine individuals, most notably Malaysian influencer Riya Ulan. The offenders overlaid the AI-generated avatar’s face—featuring an artificially darkened skin tone—onto Riya’s body, meticulously replicating her movements, attire and background settings. This brazen act of identity theft compounded the original exploitation, converting real creative work into material for deceptive accounts.

Riya’s involvement underscored the individual harm at the core of such practices. When she found her videos were appropriated and reused, she communicated her frustration to the BBC: “I was frustrated. Of course my videos are widely distributed… It doesn’t mean that you can just take it and repurpose it and share it as your own.” Her distress reveals a critical gap in platform protections, where content makers need sufficient defences against their images being stolen for deceptive pages. The incident shows how artificial intelligence amplifies current abuses, enabling bad actors to expand their abuse across many creators at once.

The exploitation extended beyond basic content piracy. Accounts deliberately constructed fake identities employing exaggerated appearance traits and synthetically modified skin tones to create striking, striking visuals intended to drive engagement. These digital constructs were marketed as real personalities, featuring invented histories and personal narratives, whilst simultaneously perpetuating harmful racist tropes and fetishisation tropes. The sophistication of the deception meant many users consumed the posts thinking they were engaging with genuine individuals, not automated creations engineered to direct users to harmful commercial offerings.

• Avatar’s face superimposed over pirated content from legitimate creators
• Digitally darkened complexions and pronounced characteristics created for engagement
• Accounts presented as genuine influencers with fabricated backstories and personas

Racist Tropes and Harmful Stereotyping

The accounts identified by the BBC and Riddance researchers displayed a troubling pattern of discriminatory targeting through their strategic naming choices and messaging. Account identifiers incorporated terms such as “black”, “noir”, “dark” and “ebony”, whilst posts consistently displayed race-based terminology and fetishizing content. Many included statements such as “loves white men” and “why I need a white guy in my life”, reinforcing damaging prejudices and portraying Black women to sexualized objects. This language pattern transformed the avatars into caricatures, sustaining historical racist tropes that commodify and dehumanise Black femininity.

The graphic depiction compounded these linguistic harms. The avatars were regularly shown in exposed swimming attire and skimpy clothing, their bodies algorithmically altered into exaggerated proportions created to enhance engagement through sexualisation. Skin tones were artificially darkened to create an artificial, near monstrous appearance that bore little resemblance to genuine human diversity. This combination of exaggerated features, exposed clothing and artificially manipulated characteristics created algorithmic stereotypes that operated like modern-day minstrelsy, converting discriminatory visuals into computational material optimised for widespread distribution and financial gain.

Why AI Exacerbates the Problem

Artificial intelligence technology has fundamentally transformed the scale and sophistication of racist exploitation online. Previously, such harmful content demanded considerable effort and manual effort to produce; AI generation facilitates the development of racist caricatures, enabling bad actors to produce hundreds of convincing fake accounts at minimal cost. The technology’s capability for producing photorealistic imagery imparts unwarranted credibility to racist tropes, making them appear credible to unsuspecting users. This technological advantage turns what was once fringe exploitation into a expandable, revenue-generating enterprise.

The computational nature of social media platforms amplifies these harms dramatically. AI-generated content designed to maximise engagement—particularly content that exploits racial and sexual stereotypes—proliferates quickly through recommendation systems designed to prioritise user interaction. Platforms find it difficult to manage the vast quantity of AI-generated content, whilst the absence of required transparency markers means users are unable to differentiate authentic creators from computational entities. This establishes conditions where racist stereotypes spread without constraint, laundered through the veneer of technological innovation and framed as entertainment rather than exploitation.

Platform Accountability and Response

TikTok’s rapid action in taking down 20 accounts demonstrates a uncommon case of platform intervention, yet it emphasises the reactive approach of content moderation in the social media age. The suspensions took place only after prolonged media attention and public backlash, implying that absent external scrutiny, these accounts would have remained active unchecked. Notably, TikTok’s response was limited to just one platform, whilst several violating accounts sustained active accounts on Instagram, where they kept collecting followers and sending users to harmful third-party platforms. This piecemeal strategy highlights the insufficiency of individual platform responses to structural issues.

Meta’s claim to be investigating, combined with its failure to announce specific steps, reveals the divide between corporate responsibility messaging and actual implementation. The company’s delayed response differs markedly from the urgency required to combat swiftly spreading damaging material. Both platforms have developed policies requiring the revelation of artificial intelligence-created images, yet these standards stayed unimplemented across numerous profiles. This implementation shortfall indicates that without government oversight and juridical responsibility, platforms will persist in prioritizing engagement metrics and advertising revenue over user safety and the safeguarding of vulnerable groups from systematic exploitation and racist imagery.

Wider Implications for Online Authenticity

The widespread proliferation of unlabelled AI-generated content on major social media platforms presents fundamental challenges about digital authenticity and user trust. As artificial intelligence becomes increasingly sophisticated, distinguishing between genuine human creators and synthetic avatars becomes significantly harder for everyday people. This loss of genuineness threatens the core assumption upon which social media communities are built—the assumption that accounts represent real people sharing genuine experiences. When millions of users unknowingly engage with fabricated personas created to manipulate them, the networks’ trustworthiness as reliable platforms for interaction deteriorates significantly. The lack of required transparency measures effectively deceives users and compromises informed consent.

Beyond personal deception, the widespread deployment of artificially created black female avatars constitutes a especially pernicious exploitation of racial identity in online environments. These artificial characters weaponise damaging generalisations and hypersexualised racial tropes whilst simultaneously appropriating authentic creators’ work and intellectual property. The practice normalises the commercialisation of black female identity within algorithmic systems designed to increase user engagement and financial returns. This algorithmic discrimination operates extensively, reaching millions whilst staying largely undetected to content moderation departments. Without robust regulatory structures requiring openness, verification protocols, and meaningful consequences for violations, online platforms will continue facilitating systematic harm against underrepresented groups.

• AI transparency obligations should be legally enforceable throughout all global platforms
• Authentication systems should verify the creator’s identity ahead of earning potential and audience expansion
• Racial stereotyping in artificially created material necessitates explicit prohibition and active detection
• Platforms need to deploy live monitoring systems with penalties aligned to content seriousness