Canva’s New AI-Powered Tool Caught Swapping ‘Palestine’ for ‘Ukraine’
by Matt Growcoot · Peta PixelCanva has apologized after its new tool, Magic Layers, was caught replacing the word “Palestine” with “Ukraine.”
As reported by The Verge, the issue was raised by X user Rosie (@ros_ie9). “I put this image into Canva [sic], I press this ‘magic layers’ button, and it turned this poster that said ‘cats for Palestine’ into ‘cats for Ukraine’.”
Other X users were also able to replicate the error, each time “Palestine” was swapped for “Ukraine.” However, words like “Gaza” were unaffected. Nevertheless, The Verge calls it “one heck of a blunder.”
The controversial bug was addressed by Canva in a statement. “We became aware of an issue with our Magic Layers feature and moved quickly to investigate and fix it,” Canva spokesperson Louisa Green says. “We take reports like this very seriously, and we’re putting additional checks in place to help prevent this in the future. We’re sorry for any distress this may have caused.”
What is Magic Layers?
Magic Layers was announced by Canva last month. It is an AI-powered feature that transforms flat images into fully editable, layered designs within the Canva editor. Essentially, it isolates different objects in an image so that they can be worked on as separate layers. A pretty nifty feature.
Magic Layers is powered by the company’s proprietary Canva Design Model, an AI system built to generate and manipulate visual designs. Unlike traditional vector tracing tools that simply outline shapes in an image, Magic Layers aims to interpret a design’s structure. The system analyzes relationships between elements, identifies text areas, and separates visual components while maintaining the overall composition.
“There’s been an explosion of AI-generated content that has, until now, been a dead end,” said Cameron Adams, co-founder and chief product officer at Canva. “You’d get a finished image you couldn’t edit, refine, or make your own. We think AI should spark creation, not stop it.”
It’s clear from that quote that the tool is designed for AI images, but it works on any kind of image, including a good old-fashioned photo.
Additional reporting by Kate Garibaldi.