In recent weeks we have been hearing more and more about a new system designed to learn to communicate with theartificial intelligence; it is a interactive gamedeveloped by Googlewhich teaches the user to create effective prompts.
Let’s talk about Say what you see and, in fact, it represents a very intuitive approach to approaching the world ofinteraction between human being and generative AI.
The user view images and must try to describe them in as detailed a manner as possible. Once the operation is completed, he can read the actual prompt used with which an artificial intelligence was asked to create the very image in question.
Say what you see is Google’s game to learn how to write prompts
Say what you see (translated into Europen as “Say what you see”) is a rather intuitive game, which aims to bring users closer without computer skills to the world of prompts.
The term prompt is used to indicate those texts written by human beings for interact with generative artificial intelligence. Think in this sense of the classic requests that are sent to chatbots like the one ChatGPT.
The game Say what you see is developed in stages and to proceed it is necessary get a minimum score.
Very simply the system proposes an image and asks the user to describe it, the player, therefore, must literally “say what you see”, translating the perception of an image into text. A process very similar to the one that leads to send written requests to the tools of generative artificial intelligence.
Come si gioca a Say what you see
Say what you see is meant for reward the user’s attention and its ability to return the greatest possible number of details in an orderly form.
To give a concrete example it is possible to cite the image used for the first level of the game: a photo representing a delicious Ramen dish.
The screen is divided into two parts: the Photo is arranged on the left side, while on the right side appears a empty boxin which the textual description of what you are observing.
Writing as a response “a plate of ramen” you pass the level. But you don’t get the highest possible score. The correct prompt in fact it is much more detailed and it can be translated like this: “A close up photo of a bowl of ramen with eggs and onion cut into rings”.
As you progress from level to level, the challenge becomes more and more complicated. And the comparison between the answers of the user ei prompt originali helps you learn the most effective strategies to communicate correctly with a generative artificial intelligence.
Say what you see is only the latest investment made by Google in support of AI development. The Mountain View giant continues to work on the experimental chatbot Google Bard. And at the same time the development of LLM Gemini family.
To know more: Artificial Intelligence: what it is and what it can do for us