Dylan Field, CEO of Figma – a collaborative web design application, has temporarily suspended the app’s newly-introduced artificial intelligence (AI) features after claims the tool reproduces designs closely resembling Apple’s Weather app. The allegations have prompted concerns about potential legal issues for users if Figma’s training data contains copyrighted content.
Figma recently showcased these new AI-powered features at its Config conference. The program’s AI tools include helping to visualize ideas, automating tasks, adjusting copy length, tone and language within designs, and generating relevant textual content in design mockups.
The controversy surrounding Figma’s AI capabilities initiated after Andy Allen, the founder of NotBoring Software, noticed that Figma’s AI tool continually generated designs strikingly similar to Apple’s Weather app. Allen accused Figma of heavily training its AI models on existing applications.
Field assumed responsibility for the issue, suggesting that this problem likely stemmed from Figma’s use of widely available AI models rather than being an intentional data training outcome. However, this raises questions regarding whether Figma should take responsibility for the outputs of its AI tools, especially given potential legal risks that its users could face for unintentionally publishing copyright content created via Figma.
Field announced that he ordered his team to disable the Make Design feature temporarily until they can stand behind its output confidently. Figma will only reintroduce this feature after successfully conducting a full QA pass on the core design system.
Still, Figma’s AI tools have garnered widespread criticism. Some critics argue that despite these tools potentially democratizing web design skills by lowering the barrier to entry, they could also propagate the mass production of low-quality, unoriginal apps.
The alleged infringement case involving Figma’s AI design feature has sparked further considerations about copyright and intellectual property rights in the AI era. An exposé by cognitive scientist Dr. Gary Marcus and concept artist Reid Southen recently uncovered the ability of AI models to generate images closely mirroring copyrighted content.
Legal challenges to AI companies using copyrighted materials in AI training data without explicit permission are on the rise. They have created uncertainty for both AI developers and content creators, whose copyright protections hang in the balance.