Amazon has upgraded its AI video model, Nova Reel, with the ability to generate videos up to two minutes in length.
Nova Reel, announced in December 2024, was Amazon’s first foray into the generative video space. It competes with models from OpenAI, Google, and others in what’s fast becoming a crowded market.
The latest Nova Reel, Nova Reel 1.1, can generate “multi-shot” videos with “consistent style” across shots, explained AWS Developer Advocate Elizabeth Fuentes in a blog post. Users can provide a prompt up to 4,000 characters long to generate up to a two-minute video composed of six-second shots.
Nova Reel 1.1 also introduces a new mode called “Multishot Manual.” In this mode, the model can reference an image along with a prompt to offer more control over a video shot’s composition. According to Fuentes, given a 1280×720-resolution image and 512-maximum-character prompt, Multishot Manual can generate videos containing up to 20 shots.
Nova Reel is only available through AWS platforms and services including Bedrock, Amazon’s AI development suite, and customers must request special access. As with most generative AI systems, there are questions as to whether Reel was developed using ethically sound methods.
Video-generating models are trained on a vast number of examples of videos to “learn” the patterns in these videos to generate new clips. Some companies train models on copyrighted videos without obtaining permission from the owners or creators, and, when these models “regurgitate” copyrighted stills, it exposes users of the models to IP lawsuits.
Amazon hasn’t revealed the source of Reel’s training data, nor has it provided an explicit way for creators possibly feeding the models’ datasets with their videos to opt out. The company has, however, said that it’ll protect any AWS customers accused of violating copyright with media generated by its models, in keeping with its indemnification policy.
Add comment