Amazon has introduced an AI-powered video generator, currently available only for advertisers, that showcases products in a new way. At its Accelerate conference today, Amazon unveiled the Video Generator, which transforms a single product image into engaging video presentations after processing. The tool can create “custom,” AI-generated videos that effectively highlight a product's features at no extra charge.
“Videos created by the Video Generator utilize Amazon’s unique retail insights to vividly animate a product story that resonates with customers," Amazon stated in a blog post.
In a statement, Jay Richman, VP of Amazon Ads, explained that Video Generator is currently in beta for select U.S. advertisers, particularly within Sponsored Brands campaigns. He added that the tool will be refined over the coming months in preparation for a broader release. “Video Generator represents another significant innovation that harnesses generative AI to foster creativity and deliver added value for both advertisers and consumers,” Richman noted. “We are committed to advancing generative AI applications that enable advertisers to create visually striking, high-performing ads.”
Also announced today is a related feature, Live Image, which generates brief, animated GIFs from still images. Like Video Generator, Live Image is currently in limited beta and is part of Amazon's suite designed for marketers, known as Image Generator.
While Amazon did not provide sample videos or extensive technical specifications about Video Generator or Live Image, a representative did answer several questions via email. The generated clips are 6 to 9 seconds long, presented in 720p resolution (24 frames per second), and are automatically produced based on a chosen product image. Each video takes up to five minutes to create, with users receiving four variations to select from.
The basic structure of Video Generator outputs includes two scenes with headlines (e.g., “Nourishing Lavender Moisturizer”), background music, and a call to action. Users can customize the font, soundtrack, and the brand’s logo displayed in the corner of the frame.
Amazon’s entrance into generative video comes as competitors also roll out similar technologies. Startups such as Runway and Luma have launched generative video APIs this month. Google is integrating its main video model, Veo, into YouTube Shorts, while Adobe has promised that video generation will be part of its Creative Suite by year-end. Additionally, OpenAI is expected to release its technology, Sora, this fall.
As with all generative AI tools, these innovations carry certain risks. Video-generating models are trained on a broad array of video content to identify patterns and create new clips. Some companies might train their models using copyrighted videos without the necessary permission, which could result in IP lawsuits if these models reproduce copyrighted material.
Amazon has indicated that it will provide protection for customers facing copyright claims related to media generated by its models, in accordance with its indemnification policy. We have reached out to Amazon to confirm whether Video Generator and Live Image fall under this policy.
Regardless of the legal outcomes, it's becoming increasingly clear that generative AI has the potential to revolutionize the film and television industry. A 2024 study commissioned by the Animation Guild, a union representing Hollywood animators and cartoonists, forecasts that over 100,000 entertainment jobs in the U.S. will be impacted by generative AI by 2026.
Updated 9/19 at 4 p.m. Pacific: Added technical details about Video Generator.
Ranking
