Hey everyone !
I wrote a post this summer about the new text-to-image AIs and their potential for creating positive impact. You can read it below if you want:
Now, some of these AIs are becoming available to the public and I have been playing around with them myself :)
For example the cover photo for this post was generated when I asked the AI for "green and gold, cyberpunk leaf".
In particular, I've had a good time with the Midjourney AI. It will give you access to a Discord where you can write your prompts and it will generate images for you. Then you can ask it for variations of these images or upscaling them.
I'll start by just showing some examples of stuff I created, along with the prompts I gave:
radioactive guinea pig
gong fu cha
It's the name of the traditional Chinese tea ceremony.
gold and green, company logo, planet
I did tons of variations of this to help me rework the branding of a digital agency I cofounded: Engines of Impact. I'm going to use this as our new logo !
Mbappe laughing at climate activists
Ok this last one is just a quick reference to the football club Paris Saint-Germain that took a private jet to go from Paris to Nantes for a game recently. Paris-Nantes is just a 2 hour train ride so a journalist asked the football player Mbappe and his coach if they would take trains more to reduce their carbon footprint. They laughed it off and made fun of the journalist by saying they would take sand yachts from now on (reference below).
So I'll reiterate the question I asked in the first post I wrote about this. How might we use these text-to-image AIs to create positive social and environmental impact.
The fact that I could rework my branding quickly is great because it will help me reach out better to impact organisations that we can help with technology.
With the Mbappe example, I think it shows that we can quickly create illustrations about current events in order to shift perceptions.
I'm sure we can think of better things still and I'd love to see what you come up with. Just go to the Midjourney website, log in with Discord and post your results in the comments below :)
Top comments (7)
I'm also unsure about the ecological impacts of these models. So many people are using them to create images right now that it could be interesting to figure out the ecological impacts of all the servers running.
Interesting read. Thanks for sharing. Any idea if these are copyright protected or free to use for any purposes ?
Glad you like it !
Reading their terms of agreement, it seems to me like they're free to use for any purpose if you have a paid plan. If you have a free plan you may not be able to resell your productions.
Is this Free or paid?
Hey Melissa !
All of the image generation AIs have free plans where you can create a bunch of images every month. I think they're quite generous if you just want to play with it. I've also used some of the paid plans for projects and I think they are quite cheap. For example OpenAI's Dall-E AI has an API where you pay for each image generation and it's very cheap :)
There's also Stable Diffusion which is an open-sourced AI model for image generation. If you can self-host it it's by far the cheapest, and you can easily train it on more images to have something that's customised for your needs.
If you're interested in that last part, I recommend checking out replicate.com/ for easy hosting of your own AI model. And I made a quick tutorial and simple process if you want to train that AI on images of yourself for example. It's here if needed: askrayan.super.site/
I am shocked at what beautiful pictures can be created.
How can I do these and what AI app is in need and free also š