×
Figma’s AI Design Tool Temporarily Removed After Inadvertently Including Copyrighted Assets
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

Figma’s AI design tool mistakenly included copyrighted assets, leading to its temporary removal for additional quality control measures.

Key issues with Figma’s AI tool: The “Make Designs” feature, launched as part of Figma’s Config event, generated outputs suspiciously similar to Apple’s weather app when given certain prompts:

  • A user discovered that asking the AI to design a weather app would produce designs closely resembling Apple’s own app, potentially exposing users to legal issues.
  • This finding suggested that Figma may have inadvertently trained the AI model on Apple’s proprietary designs, despite CEO Dylan Field’s initial statement denying the tool was trained on Figma content or app designs.

Figma’s response and explanation: In a blog post, Figma VP of Product Design Noah Levin clarified the issues that led to the tool’s removal and outlined steps for improvement:

  • While Figma carefully reviewed the AI’s underlying design systems during development and private beta testing, additional components and example screens added just before the public launch were not vetted thoroughly enough.
  • Some of these newly added assets resembled real-world applications and appeared in the AI’s output when prompted with certain queries, leading to the similarity with Apple’s weather app.
  • Upon identifying the problematic assets, Figma removed them from the design system and disabled the “Make Designs” feature to develop an improved quality assurance process before re-enabling the tool.

Design systems powering the AI tool: Levin provided insight into the extensive design systems commissioned by Figma to enable the AI model’s composition capabilities:

  • Two comprehensive design systems, one for mobile and one for desktop, were created with hundreds of components and examples showcasing various ways to assemble them, guiding the AI’s output.
  • Metadata from these hand-crafted components and examples is fed into the AI model alongside user prompts, allowing the model to assemble a subset of components inspired by the examples into fully parameterized designs.
  • Amazon’s Titan diffusion model then generates the necessary images for the design, effectively helping users identify, arrange, fill out, and theme small composable templates from the design system as a starting point.

Looking ahead: While Figma has temporarily pulled the “Make Designs” feature, other AI tools announced at Config, such as text generation for designs, remain available:

  • The company has set an August 15th deadline for users to opt in or out of allowing their data to be used for training potential future AI models.
  • Default settings for data usage vary based on user plan, with Starter and Professional plans opted in and Organization and Enterprise plans opted out.
  • Figma is working on improving its quality assurance process for AI-generated designs before re-launching the “Make Designs” tool, though no specific timeline has been provided.

Broader implications: The temporary removal of Figma’s AI design tool highlights the challenges and risks associated with integrating generative AI into creative workflows:

  • As AI models become more advanced and capable of producing high-quality outputs, the potential for unintentional copyright infringement and legal issues increases.
  • Companies developing AI-powered tools must strike a balance between providing users with powerful creative aids and ensuring that the training data and outputs do not infringe upon existing intellectual property.
  • Thorough vetting processes and quality control measures are crucial to prevent the inclusion of copyrighted assets or the generation of outputs that too closely resemble real-world designs, which could expose both the tool’s creators and users to legal consequences.
  • Transparency regarding AI training data and the establishment of clear guidelines for user data usage will be essential as more companies integrate generative AI into their products and services.
Figma explains how its AI tool ripped off Apple’s design

Recent News

Big Tech and AI startups are starting to choose leaders by lottery — why that’s a good thing

As tech companies and AI startups adopt sortition, the ancient practice of random selection gains traction as a modern tool for addressing trust and representation in decision-making.

ChatGPT Advanced Voice arrives on Mac and Windows

OpenAI's Advanced Voice mode brings conversational AI to desktop computers, enabling hands-free interaction with ChatGPT while users work on other tasks.

This new AI model aims to reduce unnecessary cancer treatments

The AI-powered diagnostic test aims to provide more accurate risk assessments for breast cancer patients, potentially reducing unnecessary aggressive treatments.