During its AWS re:Invent event, Amazon introduced a generative AI that can be used to create and/or edit images. The new AI is called “Titan Image Generator” and is part of Amazon’s AI platform Bedrock, which Amazon launched in April.
Titan Image Generator is a text-to-image AI that can create images in a similar way to OpenAI’s Dall-e and Stability AI’s Stable Diffusion. It can also edit images in a similar way as can be done with various AI functions in Photoshop. The AI can also create image captions and other relevant information about the images created, which can be used if the images are to be used for posts on social media, for example. Amazon writes about Titan Image Generator:
The model can understand complex prompts and generate relevant images with accurate object composition and limited distortions, reducing the generation of harmful content and mitigating the spread of misinformation. Customers can use the model in the Amazon Bedrock console either by submitting a natural language prompt to generate an image or by uploading an image for automatic editing, before configuring the dimensions and specifying the number of variations the model should generate.
Amazon’s new text-to-image AI will also insert invisible watermarks in the images created or edited with the AI, a measure to prevent the images from being used, for example, in disinformation purposes.
Amazon’s Titan Image Generator is already available in a preview version, and Amazon plans to roll out a public version for AWS customers in the coming months. In the video clip below, Amazon talks more about its new text-to-image AI.