This Google Game Secretly Teaches You Perfect AI Image Prompts

Why It Matters

It equips a wide audience with the core skill of prompt engineering, boosting the effectiveness of AI‑generated imagery and accelerating its integration into creative and commercial workflows.

Summary

The video spotlights Google’s new interactive experiment, “Say What You See,” a gamified tool that trains users to craft precise AI image prompts. By presenting an AI‑generated picture and challenging players to describe it in fewer than 120 characters, the game turns prompt engineering into a quick, repeatable exercise rather than a tedious tutorial.

The mechanics are straightforward: after submitting a description, the system regenerates an image and assigns a “math score” based on visual similarity to the original. Players receive three attempts per round, and each successive level tightens the criteria—demanding sharper attention to subject, lighting, style, and contextual cues. This iterative feedback loop forces participants to refine their language, honing the balance between brevity and specificity that modern text‑to‑image models require.

Developed in partnership with artist Jack Wilde, the experience is pitched at creators, designers, students, and anyone new to generative AI. It explicitly references popular models such as Google’s Imagen and OpenAI’s DALL·E, positioning the game as a low‑stakes sandbox for mastering the “prompt‑craft” skill set that underpins high‑quality AI‑generated visuals.

The broader implication is a democratization of prompt engineering—a skill that has quickly become a bottleneck in creative workflows. By embedding learning in a playful format, Google lowers the entry barrier, potentially accelerating adoption of generative tools across industries and improving the overall quality of AI‑driven visual content.

Original Description

Google’s “Say What You See” turns AI prompting into a game—helping you master image prompts without tutorials or homework.