Google’s Say What You See Game Helps You Learn Image Prompting

Google has launched a game that helps users learn “the art of image suggestion” to create images using generative AI. The game called “Say What You See” was developed by the Google Arts & Culture team. It's a fun way for kids to learn about image prompts and how image generation works at “https://artsandculture.google.com/experiment/say-what-you-see.

This is how the game works

As the name suggests, it shows users an AI-generated image on the left and a text prompt on the right. Users need to create an appropriate prompt for the image and enter it into the text field. However, the prompt must be 120 characters long. Based on the prompt, Google AI will generate an image, compare the two images, and suggest a match percentage.

The goal is to get a higher match percentage

A higher match percentage indicates that the user entered a better prompt. Additionally, there are multiple levels in the game. The first level requires users to get a 50 percent match, while the second and third levels require users to enter prompts carefully enough to get a 60 percent match. and 70 percent, respectively. Depending on the levels, users can get the necessary level on the first attempt or take up to three chances per frame.

With each answer users enter, the game also provides a suggestion that helps users improve the prompts, for example by adding more details about the nature of the image, topic, or context. While there is no special prize for completing all levels, the game teaches users about image prompts, what to include, and how to generate better images in shorter descriptions.

You can follow Smartprix on Twitter, Facebook, Instagram and Google News. Visit smartprix.com for the latest news, reviews and technical guides.