Social media influencers say artificial intelligence tools on platforms like TikTok and Instagram are altering their racial appearance in posted content [1, 2].

This trend raises concerns about the stability of digital identity and the potential for AI to erase or distort ethnic characteristics without user consent. As AI-driven filters become more sophisticated, the line between intentional creative expression and involuntary modification is blurring.

One female influencer said her race was changed by AI on social media and said she is not alone in this experience [1, 2]. These alterations occur via generative filters that can modify skin tone, facial structures, and other ethnic markers.

Some creators use these AI tools to experiment with or monetize their visual identity [3]. This process allows influencers to overhaul their appearance or play with different identities to attract specific audiences, a practice that has become more common as AI accessibility increases [3].

However, the reports of involuntary swaps suggest a different phenomenon. While some users intentionally manipulate their image, others describe a process where the AI makes these changes autonomously. This discrepancy highlights a tension between AI as a tool for self-expression and AI as a source of unintended distortion [1, 3].

Platforms like TikTok and Instagram have integrated various AI-powered effects that analyze user data to apply filters in real time. The technology relies on datasets that may not always accurately represent the diversity of human features, which can lead to the perceived swapping of racial traits [1, 2].

One influencer said her race was changed by AI on social media.

The emergence of AI-driven race swapping underscores a critical gap in how generative algorithms perceive and render human ethnicity. When AI involuntarily alters a person's racial markers, it suggests that the underlying models may be applying biased or generalized templates to diverse faces. This could lead to a digital environment where ethnic authenticity is compromised by the software used to share it.