San Francisco: Canva has issued an apology after users flagged an issue with its AI-powered feature that replaced the word “Palestine” with “Ukraine” in designs, sparking widespread backlash and raising fresh concerns about bias in generative tools.
The controversy centres on Canva’s “Magic Layers” feature — part of its expanding AI toolkit — which is designed to break down flat images into editable components. Instead of simply separating text and visuals, the tool unexpectedly altered written content, triggering debate over reliability and moderation in AI systems.
Issue goes viral after user report
The problem first came to light when an X (formerly Twitter) user demonstrated how the phrase “Cats for Palestine” was automatically changed to “Cats for Ukraine” while using the Magic Layers feature. The post quickly went viral, with many questioning why an editing tool would modify text without user input.
Users experimenting with the feature reported that the issue appeared to be specific to the word “Palestine”, while other related terms such as “Gaza” remained unaffected. Several claimed they were able to replicate the behaviour before the company rolled out a fix.
Canva responds with apology and fix
Responding to the backlash, Canva acknowledged the issue and said it had been resolved. In a statement to The Verge, a company spokesperson said it had acted quickly to investigate and correct the problem.
“We became aware of an issue with our Magic Layers feature and moved quickly to investigate and fix it,” the spokesperson said, adding that additional safeguards are being introduced to prevent similar incidents.
The company also apologised to users, stating it takes such reports seriously and regrets any distress caused. However, it did not provide a detailed explanation of how the error occurred.
Questions over AI behaviour and bias
The lack of clarity around the cause has become a key point of concern. Experts note that AI systems can behave unpredictably due to the vast datasets they are trained on, as well as internal moderation layers that may influence outputs.
In this case, the unexpected substitution of a politically sensitive term has intensified scrutiny. Critics argue that even isolated incidents can reflect deeper issues in how AI models handle language, especially around geopolitically charged topics.
Timing adds to pressure on Canva
The incident comes at a crucial time for Canva, which has been aggressively expanding its AI offerings to compete with rivals like Adobe. Magic Layers is part of a broader push to position the platform at the forefront of AI-assisted design.
Instead of highlighting innovation, the feature has now drawn attention for an embarrassing error, potentially impacting user trust in its AI capabilities.
Wider pattern across AI platforms
This is not the first time AI tools have faced allegations of bias involving politically sensitive prompts. Several tech platforms have previously been criticised for problematic outputs linked to global conflicts and terminology.
Each such incident adds to growing concerns that generative AI systems may inherit hidden biases from training data or exhibit unintended behaviour due to complex rule-based filtering.
Conclusion
While Canva has fixed the issue and apologised, the episode underscores ongoing challenges in developing reliable AI tools. As companies continue to integrate AI into creative workflows, ensuring transparency and consistency will be critical to maintaining user trust.