Support for Google I/O
Improving data quality for generative AI images
What I did
Project
Role: Data quality (images)
Company: Google I/O
Problem
For Google’s flagship event Google I/O 2023, several teams put together a game called I/O Flip using generative AI tools that were going to be featured at the event.
“I/O FLIP is an AI-designed take on a classic card game, powered by Google, and created to inspire developers to experiment with what is possible with Google’s new generative AI technologies. Thousands of custom character images were pre-generated with DreamBooth on Muse and their descriptions were written using the PaLM API. The game’s UI and backend were built in Flutter and Dart, a suite of Firebase tools were used for hosting, and sharing, and Cloud Run was used to help scale.”
But the technology was new and the images rendered weren’t always on-brand.
Solution
The teams asked a handful of people, including me, to help parse through the images, select the high quality ones, and provide feedback to the team on what wasn’t useable to better train technology.
Impact
The game was featured at Google I/O, Google I/O Connect, and the Google for Developers blog, garnering over 1M views.
Learn about the game
Tab through the images below to see the game or read more about how it was made here.