Amazon has released an image generator – joining the ranks of many, many other tech giants and startups that have already done so.
During a keynote at the re:Invent conference this morning, Amazon debuted the Titan Image Generator, which is now available in preview for AWS customers. The Titan Image Generator, a part of Amazon’s Titan family of generative AI models, can create new images given a text description or customize existing images.
“You can use the model to easily exchange a background with the background of a rainforest (for example),” said Swami Sivasubramanian on stage. “(And you can) use the model to seamlessly swap backgrounds to create lifestyle images, all while maintaining the main subject of the image and making a few more choices.”
Amazon says the Titan Image Generator is trained on a “diverse set of data sets” in a “broad range of domains” and includes built-in mitigations for toxicity and bias. The company declined to say where the data sets came from, however — and whether it got permission from all the creators of the images it uses to train the Titan Image Generator.
Sivasubramanian claimed on stage, however, that Amazon will protect customers accused of infringing copyright on images created using the Titan Image Generator.
“We’re very selective about how we train our models and the data we use to do it,” he said.