Canva, the graphic-design platform, unveiled an AI-powered feature to generate images from text prompts.
Here’s a #CanvaTip to turn your ideas into an image! 🖼️https://t.co/VRaamvQD8L
Go to the More tab and select Text to Image. Describe the image you want to see and then choose your style of photo. pic.twitter.com/0NmNFmwWRz
— Canva (@canva) November 8, 2022
Text to Image is based on the open-source text-to-image model Stable Diffusion. The developers added several extra safety filters and a configurable interface.
To use the feature, describe the desired image and choose one of the styles. The model will output a grid of four images and add them to the designer’s canvas.
Users can also report images containing violence, nudity, hate speech, and ‘biased or stereotypical’ content.
The feature is available for all plans, including the free tier. Users can create up to 100 images per day.
«We aim to present this technology to the community because it is a developing field», said Canva co-founder and Chief Product Officer Cameron Adams in an interview with The Verge.
According to him, the tool is already being used in many areas.
«One of my favourite ways is how students use it to visualise their stories […]. We have also seen how it [the tool] creates images for presentations, flyers and t-shirts that they [users] can print through Canva», Adams said.
He also noted that the company is examining copyright questions around AI-generated images:
«To some extent, AI-produced content may be considered fair use, and this will vary across countries. We are closely watching the situation».
Canva began beta testing the tool in September 2022. The feature is now available to all 100 million users of the platform.
In November, the PhotoRoom learned to change the background on images with AI.
In October Shutterstock expanded its partnership with OpenAI to sell stock images generated by the DALL-E 2 model.
Subscribe to ForkLog news on Telegram: ForkLog AI — all the news from the AI world!