Amazon has introduced an update for its generative AI model Nova Reel, which can now create videos up to two minutes in length. Nova Reel, first announced in December, marked Amazon’s initial attempt to enter the generative video field, where models from OpenAI, Google, and other companies are already actively competing.
The new Nova Reel 1.1 version adds the ability to generate “multi-shot” videos with a “consistent style” between frames. Users can enter a prompt of up to four thousand characters to create a video consisting of six-second shots. A new “Multishot Manual” mode has also been introduced, allowing users to combine images with prompts for greater control over frame composition. In this mode, you can generate videos with up to twenty frames, using images with a resolution of 1280×720 pixels and a prompt of up to 512 characters.
Nova Reel is available only through AWS platforms and services, including Bedrock, Amazon’s AI development toolkit. Users must request access, but AWS automatically approves these requests. As with most generative AI systems, there are questions about the ethics of the model’s training methods.
Amazon does not disclose the data sources used to train Nova Reel and does not provide content creators with the ability to opt out of having their materials used for model training. However, the company has stated that it will protect AWS users accused of copyright infringement related to media generated by its models, in accordance with its customer protection policy.
