A messaging app that uses AI-generated images as the only form of communication
Image credit: OpenAI's DALL-E 2
Darko SavicJun 13, 2022
Please leave the feedback on this idea
Is it original or innovative?
Is it feasible?
Is it targeting an unsolved problem?
Is it concisely described?
Bounty for the best solution
Provide a bounty for the best solution
Bounties attract serious brainpower to the challenge.
A messaging app where AI converts your sentence to a single image that perfectly conveys your message. Communication is exclusively visual.
A new, fun way to communicate.
Now a picture is literally worth a thousand words.
A message might be cheesy in words but make a lot of sense as an image.
How it works
You write a message to your friend, just as you normally do. When you hit send, you get a few images that are a representation of the message you just wrote. The AI conveys your words with one image (a picture worth a thousand words). Pick the one that best represents your message and send it.
Seeing the image, your friend tries to figure out what you are saying. They then write a reply and the AI converts it to an image again. This goes back and forth. It's a new, fun way to communicate.
The technology is ready
The technology that makes this idea possible is called Dall-e from Openai. It's a new AI system that generates realistic images or art, from a description that you provide it:
Google is also working in the space with it's Imagen AI:
Photos with actual people
In this AI-powered imaging messenger you can add a few photos of each person in your contact list. The AI can then use your friends' faces and place them in the image according to your description. When appropriate the AI can create realistic photos of you and your friends, as if they actually happened.
It won't be long before AI can create short movies based on your description. The movies can feature you and your friends. The good ones can potentially go viral in your circles.
Effective human communication through images would take an upgrade in existing technology
Povilas SJun 21, 2022
Existing image-from-text generators that you linked are oriented to generate images in a descriptive, "noun" manner. With the generated image they answer the question "what is there"?
Language goes way beyond that. People communicate a lot about how they feel. Imagine two people trying to clarify their relationship - I feel this and that because you did certain things, etc. I guess current image-generating systems would not be effective in conveying this. It would take a succession of images, like a movie, and still, the final result would be too prone to interpretations. The asking for clarification might not seem very clear as well and the misunderstandings would build up from there.
Some words, concepts, and meanings are hard to convey through visual objects. We understand the meaning conveyed in words, because we relate many experiences to certain words and paint a general "picture" of what it means in our heads, but that picture is personal and complex, painted through years of life experience. However, in the case of words, at least we have a generalized system of symbols to "fix" the meaning in place. With generated images that fixation would be lost and it would be prone to interpretations way more than words are.