In this blog John, one of our software developers, takes a look at whether AI art generators could become a tool in a creatives kit, and whether it'll see the light of day in software of the future.
Dall-E 2 is OpenAI’s latest AI model for generating images from natural language prompts. It allows a user to enter a prompt and then generates a series of images from your input.
For example: ‘An astronaut riding a horse’
You can further augment your prompts by adding a particular artist's name or artistic medium.
For example: ‘An astronaut riding a horse as a pencil drawing’
I’ve recently gained access to Dall-E 2 (after joining their waiting list earlier in the year). It became clear that its general functionality is hugely powerful but after running a few ‘fun’ prompts, you start to realise that without real purpose it quickly becomes a bit of a gimmick.
Gimmicks aside, it's also clear that Dall-E 2 being as powerful as it is means it could have real potential to help aid creative professionals. Certain creative tasks could easily be automated using Dall-E 2. In particular, tasks that have clearly defined parameters could work really well.
For example:
I’ve taken some of the past blog post themes on the Newicon website and run them as prompts to see how it would handle generating blog post images.
Technologies that will define the future of the IoT - Prompt used: ‘Technologies that will define the future’
The result is two perfectly acceptable images. Dall-E 2 picked up on the key concepts of the prompt and made a futuristic style image that would be perfect as a blog featured image.
Joining Bristol24/7’s Better Business Network - Prompt used: ‘Joining Business Network’
These images are exactly what you’d expect to see if you searched a stock image library for this prompt. I enjoy the fact they are basically the same concept but with slight variations (a different viewing angle and the figures are differently coloured in the second image).
As an example of generating on-the-fly stock photography, I used the prompt ‘photograph of a cat holding a letter N’.
Theoretical video for YouTube titled ‘Why cats love Newicon’
If you’re more of a dog person, then Dall-E 2 has got you covered. I used the prompt ‘photograph of a puppy holding a letter N’.
Theoretical video for YouTube titled ‘Why dogs love Newicon’
I can see integration with Dall-E 2 (and/or its successors) becoming an everyday element in software.
Take for example a blogging platform that automatically generates a featured image using your post title. It could suggest a few options and you pick the desired image. My above examples prove that it's completely capable of this currently.
Dall-E 2 (and/or its successors) could quickly become a useful tool in any graphic designer's arsenal. Instead of searching for the exact stock photograph you need, you’ll be able to generate it with a few clicks. This could be integrated into popular software like Adobe Photoshop and simply offered as another menu item, making your creative process even more seamless. While I don’t think it’s time to be retraining if you are a graphic designer (or similar), it’s certainly time to sit up and take notice.
Want to talk to us about how to integrate AI into your next software project? Set up a call with Mark today.
Subscribe to get our best content. No spam, ever. Unsubscribe at any time.
Send us a message for more information about how we can help you