YouTube’s AI deepfake monitoring expands to celebrities, spotlighting corporate control over digital likeness rights and labor exploitation in the attention economy

Indigenous frameworks treat likeness as inseparable from community and land, making AI deepfakes a violation of collective rights rather than a personal grievance. Traditional knowledge systems often prohibit the replication of sacred or ancestral imagery, a principle violated by AI-generated likenesses of cultural figures. The focus on celebrity removal requests ignores how AI deepfakes erase Indigenous narratives by substituting corporate-controlled replicas for living traditions.

The commodification of likeness traces back to 19th-century copyright laws designed to protect elite performers, not marginalized groups, while enabling corporate exploitation. Historical precedents like blackface minstrelsy show how digital deepfakes are a modern iteration of dehumanizing representation economies. The current legal framework mirrors 20th-century struggles over unauthorized biopics, where courts prioritized corporate profit over individual dignity.

In Japan, the concept of 'seken' (public reputation) has led to legal battles over unauthorized AI likenesses, blending Western individual rights with communal accountability. African digital rights activists argue that AI deepfakes disproportionately target women and LGBTQ+ creators in contexts where legal recourse is weak. South Korea’s 'right to one’s own image' laws offer a model for balancing celebrity protections with collective cultural rights, though enforcement remains uneven.

AI deepfake detection relies on flawed datasets that underrepresent non-Western faces, leading to higher false positives for marginalized groups. Research shows that current likeness detection systems often conflate stylistic mimicry with malicious deepfakes, risking censorship of creative expression. The scientific literature highlights the need for adversarial testing to prevent bias in enforcement, yet YouTube’s system lacks transparency on its training data or error rates.

Artists like Hito Steyerl critique AI likeness tools as extensions of surveillance capitalism, where creativity is reduced to monetizable data points. Spiritual traditions from Sufism to Indigenous animism warn against the hubris of replicating the human form, framing AI deepfakes as a violation of sacred boundaries. The erasure of artistic intent in AI-generated likenesses reflects a broader devaluation of human expression in digital economies.

If unchecked, AI likeness control could lead to a 'digital feudalism' where corporations own the rights to all human expressions, as seen in proposals for 'personality rights' in AI training data. Scenario modeling suggests that without global treaties, deepfake removal will become a luxury service, exacerbating inequalities between Global North elites and Global South creators. The rise of 'synthetic celebrities' may redefine labor, with AI-generated personas replacing human creators entirely.

Non-celebrity creators—especially sex workers, dancers, and gig workers—face the highest risks of deepfake exploitation due to lack of legal protections and platform bias. Black and brown performers report higher rates of unauthorized AI likenesses in music and film, reflecting historical patterns of dehumanization. Disabled creators are often targeted by deepfakes that mock or exploit their identities, yet their cases are deprioritized in content moderation systems.