TikTok mistakenly posted a link to an internal version of its new artificial intelligence avatar tool, which has no guardrails and lets users create videos expressing almost anything. CNN (CNN) first discovered the problem and allowed the outlet to produce videos containing phrases such as Hitler quotes and messages telling people to drink bleach. TikTok has since removed that version of the tool, while the version TikTok intended to launch remains available.
TikTok’s Symphony Digital Avatars, launched earlier this week, allow businesses to create ads using the likenesses of paid actors. It also uses AI-powered voiceovers, letting advertisers input scripts to have avatars say what they want within TikTok’s guidelines. Although only users with a TikTok Ads Manager account can access the tool, the version spotted by CNN allows anyone with a personal account to try it.
in a statement edgeTikTok spokesperson Laura Perez said TikTok had fixed a “technical error” that “allowed a very small number of users to create content using an internal beta version of the tool for several days.”
When CNN discovered this internal tool, it allowed the outlet to generate videos reciting Osama bin Laden’s “Letter to America,” a white supremacy slogan and a message telling people to be in the wrong Date voting videos. None of the videos produced by CNN show the watermark that the video was generated by artificial intelligence, which is exactly what the correct version of TikTok’s digital avatar of the symphony does.
CNN did not post the videos it produced to TikTok, but Perez noted that if it had, the content “would have been rejected for violating our policies.” Although TikTok has since removed that version of the tool, it has raised questions about whether people will find other ways to abuse avatar creators and whether TikTok is prepared for this.
