Canva’s AI quietly swapped ‘Palestine’ for ‘Ukraine’ in user designs

I suddenly feel so much better about every embarrassing typo I’ve ever made. Turns out, AI can make your designs politically incorrect without even asking.

One of Canva’s newer AI features—Magic Layers—was caught doing something it absolutely shouldn’t: silently replacing the word “Palestine” in user designs. The feature is supposed to break flat images into separate editable layers, not rewrite your text. But X user @ros_ie9 found that when they used it on a design reading “cats for Palestine,” the output came back as “cats for Ukraine.”

This wasn’t a random glitch. The same user noted that related terms like “Gaza” were left untouched. Only “Palestine” got swapped. That’s a pretty specific substitution for something that claims to be a neutral, technical tool.

Canva acknowledged the issue and said it has since resolved it, with steps to prevent a repeat. That’s the corporate line. But honestly, this feels like a textbook example of an AI model being trained on—or post-processed with—some kind of content filter that went rogue. It’s not hard to imagine a moderation layer deciding “Palestine” is problematic and auto-replacing it with a “safer” alternative. Except nobody asked for that.

This is the kind of bug that makes you wonder what other silent edits are happening under the hood. If an AI can quietly swap a country name, what else is it changing without telling you? Canva’s apology is fine, but trust takes longer to rebuild than a patch rollout.