AutoDraw was created by Google Creative Lab to make it easier to create images. AutoDraw, which was released in May 2017, is a part of Google's continuing AI Experiments series, which showcases innovative uses of machine learning. It was created by Kyle Phillips, Dan Motzenbecker, and a team at Google Creative Lab with the intention of making sketching more approachable and enjoyable for everyone.

The machine learning algorithms in AutoDraw were developed using the same technology as QuickDraw, a game created by Google's Creative Lab with the goal of generating the biggest doodle dataset ever. Players were tasked by QuickDraw to sketch something while a neural network tried to identify it. The dataset for AutoDraw was this enormous collection of doodles, which effectively trained the AI to identify forms and infer what the user is attempting to draw.

The AI makes intelligent assumptions about the thing being drawn while users create sketches on AutoDraw's canvas. It then offers a variety of professional clip art made by artists that corresponds with the drawn item. This tool has the ability to convert crude sketches into more refined, expert-looking designs. As time passes, AutoDraw's guessing abilities get better, honing its suggestions in response to human input. The AI will be more likely to offer related images in the future if a user picks a suggested piece of clip art since it interprets this action as good feedback. In contrast, if a proposal is rejected, the AI interprets this as unfavourable feedback and is less likely to offer the same advice again, thereby gradually increasing its accuracy.

The tool is made to be convenient and easy to use. Users are given a streamlined art programme interface when they first use AutoDraw in a web browser. As soon as the user starts sketching, the AI's best predictions display at the top of the window, providing the user with a number of other drawings. Once the doodle is finished, it may be shared immediately across multiple social media sites , shared via a created URL, or downloaded as a PNG file. Anyone with internet connection can access AutoDraw since it can be used on any device, including phones, tablets, laptops, and desktop computers.


  • AI-Driven Guessing: AutoDraw uses machine learning to guess what the user is trying to draw, offering a selection of professional clip art based on the user's sketch​.
  • Intuitive User Interface: The tool presents a simplified art program interface, making it user-friendly and easy to navigate. As users sketch, the AI's best guesses appear along the top of the window for easy selection​.
  • Multimedia Functionality: Users can not only draw with the tool, but also add text, shapes, and fills to their image​.
  • Sharing and Downloading Options: Once a doodle is completed, users can download the image as a PNG file, share it via a generated URL, or directly share it on various social media platforms​.
  • Learning Capabilities: AutoDraw learns from user feedback to refine its guesses. If a user selects a suggested clip art, the AI will be more likely to suggest similar drawings in the future, while ignored suggestions are less likely to appear again. This way, the tool's accuracy improves over time​.


  • AutoDraw is a free tool provided by Google, with no subscription or purchase required.

Popular Use Cases

Videos (Official, Reviews, How-To's)




Find Out How How Best To Utilise AI Tools

Our newsletter comes with exclusive discounts, trials and practical insights from within the industry