summary
Tiktok launches three new AI tools for automated video creation, making it easy for brands to generate advertising content designed for the platform.
The new “Image to Video” feature converts static images, such as product photos and brand visuals, into 5-second video clips. Advertisers can upload images, add simple text prompts, and receive video options generated by some AI to use them individually or combine them with longer ads.
Video: Tiktok
There is also a “text-to-video”. This allows advertisers to create videos from just text without the need for images or templates. Tiktok said the tool will help you quickly test and refine campaign ideas before moving to full production.
advertisement
Video: Tiktok
The third tool, “Showcase Products,” blends product images with digital avatars. Advertisers can create more dynamic, UGC-style ads by holding products in their avatars, testing clothing, and viewing apps on their smartphones.
Video: Tiktok
Tiktok will work with Adobe to bring “images” directly to Adobe Express, Adobe's social media design platform. The Symphony feature will also be integrated into WPP Open, WPP's AI-driven marketing suite.
Built-in AI labeling and safety checks
All videos made with Symphony are automatically labeled as AI generation. Since May 2024, Tiktok has been part of the CAI and C2PA industry groups. According to Tiktok, all stages of content creation – going through multiple security checks, from uploading images and writing prompts to generating final video. The company has not revealed which AI models the power symphony.
recommendation
Tiktok, owned by Chinese bytedance Ltd, first launched Symphony and AI Avatars in 2024. Three new features will be rolled out globally at Tiktok's Symphony Creative Studio over the next few weeks.
Bytedance's research team was a major force in advance of generated video AI. Earlier this year, they introduced the Goku model series designed to create UGC content with Human Avatars, a technology that can power Tiktok's latest tools. More recently, Bytedance has demonstrated Seedance 1.0. This is a video model that will face off against Google's VEO 3.
