Independent game publisher Finji has leveled accusations against TikTok for employing generative artificial intelligence to change advertisements for its titles on the social media site without any approval or awareness from the company. Known for releasing popular indie titles such as Night in the Woods and Tunic, Finji learned about these apparently altered promotions after supporters notified its verified TikTok profile.
According to a report from IGN, Finji claims that a particular advertisement distributed on the platform was changed to feature a problematic portrayal of a character from one of its games, described as involving racial and sexual stereotypes. Although the studio does run ads on TikTok, it informed IGN that all AI-related features were fully disabled. Rebekah Saltsman, the CEO and co-founder, obtained images of the problematic ads from enthusiasts and subsequently reached out to TikTok for clarification.
Several Finji advertisements have surfaced on TikTok, including compilations showcasing various company games and targeted ones for specific releases, such as the title Usual June. IGN noted that the disputed AI-altered ads, which continue to appear as originating from Finji, were presented in slideshow format. While certain visuals seem close to the originals, one apparent AI-created instance observed by IGN shows the main character of Usual June wearing a bikini-style bottom, exaggerated hip and thigh proportions, and boots extending above the knees—a depiction that starkly contrasts with her actual in-game appearance, as evident from official imagery.
Regarding TikTok's reaction, IGN shared excerpts from the platform's communications addressing Finji's concerns. Initially, TikTok stated it detected no signs of AI-produced elements or slideshow styles being applied. This claim came even after Finji provided a clear image of the manipulated visual to customer support. In a follow-up interaction, TikTok recognized the proof and confirmed it was no longer challenging the occurrence, while noting that the matter had been referred to internal teams for detailed review.
TikTok offers a 'Smart Creative' tool in its advertising system, which leverages generative AI to create variations of submitted ads, prioritizing those that receive stronger engagement from viewers. Additionally, the 'Automate Creative' capabilities use AI to refine aspects like soundtracks, sound enhancements, and overall visual appeal to improve the audience's experience. Saltsman presented IGN with documentation showing that Finji had deactivated both features, a detail later verified by a TikTok representative concerning the specific ad involved.
Following several escalating discussions marked by Finji's growing irritation, TikTok eventually conceded to Saltsman that the advertisement presented major problems, encompassing improper AI application, the objectification and distortion of character designs, and potential damage to the studio's business and image. The platform then provided a partial rationale, explaining that Finji's setup incorporated a 'catalog ads format' intended to highlight advantages of integrating carousel and video components in sales-oriented campaigns. According to TikTok, this approach enables marketers to obtain improved outcomes with minimal input, though it sidestepped the concerns over inappropriate material. It appears Finji had unknowingly selected this format. When IGN sought comment, TikTok refused to provide one.
Saltsman was informed that further escalation was impossible, and the dialogue remained unsettled as of IGN's article publication. In remarks to the publication, Saltsman expressed surprise at TikTok's inadequate handling of the situation they created. She indicated a desire for both an official apology and firm commitments to prevent future incidents, while admitting she held little optimism for receiving them.