Finji Accuses TikTok of Unauthorized AI Modifications to Their Ads, Including Racist Content

Finji, known for publishing popular indie games such as Night in the Woods and Tunic, has claimed that TikTok has been modifying its advertisements using generative AI without obtaining permission. This has led to the dissemination of ads that include a racist and sexualized representation of one of Finji’s characters. The issue was initially highlighted by Rebekah Saltsman, Finji’s CEO and co-founder, on the social platform Bluesky. She shared a screenshot of a post from another brand experiencing similar issues, encouraging followers to send her screenshots of any ads that appeared uncharacteristic of Finji’s branding. In an interview, Saltsman explained that while Finji’s official TikTok account is active in promoting its games, they have ensured that AI functionalities are completely disabled. The company became aware of these unauthorized AI-generated ads through feedback from users who commented on Finji’s legitimate ads, expressing concern over what they found. Some fans provided screenshots of the altered ads, prompting Saltsman to contact TikTok support for assistance. The ads in question were originally videos promoting Finji’s games, including a focus on Usual June. However, the AI-modified versions that appeared on TikTok looked like they were posted from Finji’s official account yet featured slideshows rather than the intended video content, as noted by various comments from viewers. Finji shared these viewer-sourced screenshots which illustrated the discrepancies, with one particular image standing out due to its significant alterations. The altered image showcased an edited rendition of the official cover art, featuring the character June posed alone. The AI-modified version depicted her in a bikini bottom with exaggerated physical features, perpetuating harmful stereotypes that starkly contrast with her portrayal in the game. Further communication between Finji and TikTok customer support revealed that Finji had both the “Smart Creative” and “Automate Creative” features disabled. These TikTok options utilize generative AI to create multiple versions of ads and optimize their elements for better viewer engagement. Saltsman presented screenshots that confirmed these features were turned off on their account. Finji expressed frustration at being unable to view or modify the AI-generated ads, relying instead on user comments and Discord reports to inform them of the issue. Saltsman noted suspicions of other inappropriate AI-generated ads featuring another character, Frankie, based on user feedback, though she lacked the means to verify these modifications. Despite reaching out for support, TikTok’s representatives did not provide a satisfactory resolution. The support team initially stated that there was no evidence of AI-generated content in Finji’s ads, which contradicted the reports and evidence provided by Finji. As the dialogue continued, TikTok acknowledged the seriousness of the situation and confirmed that they would escalate the matter for further investigation. However, after several follow-ups, Finji was still left without a direct line to a senior representative and faced vague responses about the origins of the AI-generated content. In a subsequent message, TikTok attempted to explain the appearance of the AI-modified ads, suggesting that Finji’s campaign had inadvertently used a catalog ads format aimed at blending carousel and video assets. The communication emphasized potential performance improvements for advertisers, but failed to address Finji’s concerns about the offensive content. Finji expressed their dissatisfaction over TikTok’s handling of the situation, questioning the lack of action regarding the sexualized and racist representations of their work, and seeking clarity on the tracking of AI-generated ads. The publisher’s ongoing struggle to resolve this issue highlights significant concerns around unauthorized modifications and the implications of generative AI in advertising.