TikTok has added a new “AI greenscreen” capability to its platform that allows creators to transform text descriptions into images, which they can then use as video backgrounds to make content that looks even more unique. The feature is accessible via TikTok’s camera screen.

When users select the effect, they are presented with various sample prompts such as: “hidden village in the mountains,” “snorkeling in a purple ocean,” and “erupting rainbow volcano.” TikTok has also incorporated safeguards to ensure people adhere to the app’s community guidelines, which company representatives say are tantamount to the platform’s mission. For example, if a user types in text containing nudity, profanity, or violence, only abstract, unclear images with colors and themes will appear.

“At TikTok, we aim to build responsibly and equitably so our community can safely create, share, and enjoy creative and entertaining content on our app,” Vanessa Pappas, chief operating officer at TikTok, said in a statement, referring to the company’s commitment to transparency.

Text-to-image AI generators have gained traction since OpenAI released its software DALL-E last year, which, along with Google’s Imagen and Midjourney, provides highly sophisticated image generation. TikTok’s “AI greenscreen” is comparatively basic, and cannot produce the same level of photorealistic imagery or complex illustrations that look like they were drawn or painted by actual people.

However, given its user base of over one billion people, TikTok may be intentionally limiting its new offering in order to mitigate implementation costs and resources, and steer clear of issues that could arise if it were to offer a more advanced AI feature on such a large scale.