Pick a Pic is an open-source tool for gathering human commentary on AI-created visuals for scholarly AI research. Users of the app can create graphics from text and then rate them.

The study project’s code, data, and models are all publicly accessible and open-sourced. Before creating and taking part in the project, users must authenticate with Google.

The goal is to create the largest text-to-image collection with human feedback that is publicly accessible and open-sourced. The app currently employs the ProtogenX3.4 and Stable Diffusion 2.1 models.

The collected data will be made available to the public and will be updated on a regular basis. The Google TRC Program is funding this study.