By LOVELIVE Lead Creative, Garrick Middleton
The world of AI asset creation is moving so fast that this is literally the 3rd time I’ve had to re-write this blog in just a week!
AI image generation is blowing up at the minute. I believe that it’s a proper revolution in workflow – as Jensen Huang, CEO of NVIDIA, stated in his GTC keynote “We are at the iPhone moment of AI”. There’s clearly huge, huge potential for AI services to impact our processes – in the same way that Netflix impacted movie consumption.
And just to be clear, this isn’t a comment on the implications of AI using copyright-protected images to feed its learning and image generation capability. This is set to be addressed by Adobe, Shutterstock, and Getty (among others) as they prepare for their own AI tech that will be trained on their own libraries.
But I really don’t want to get into that argument right now – this is more about how I’m finding AI fun, engaging, and more importantly a powerful design tool. For instance, Version 5 of Midjourney – one of the leading AI image generation platforms – launched recently, and I have to say it’s mind-blowing. I’ve been using it mainly for generating photographs of realistic people – as I reckon if they can nail that, then they’ve done a good job.
You see, in Version 4, the people in the images were fully-paid up residents of that place we call “Uncanny Valley”. Odd teeth, strange expression, a finger or two too many per hand, a painterly look to the photo, etc.
But in V5: wow! It’s scary good – you can look into the person’s eyes and it just ‘feels’ like you’re looking at a real person, with a real life and real memories. It’s mad really.
V5 has really made AI people image creation a viable alternative to stock libraries. Not to mention that V5 has doubled the resolution of the renders it generates – further helping the process.
To show how it’s already having a real business impact for our clients, here’s a great way we’ve used AI to solve a real design issue:
A recent client acquisition of ours had the tricky problem that they needed images of people that relate to their audience BUT the subjects of the images would need to be anonymous. Using stock images wouldn’t work either, as the subject matter would break stock Ts & Cs.
This is where the power of AI came in and, especially now with V5 Midjourney, we’ve been able to create bespoke images that reflect their audience, from many-a demographic, whilst completely side-stepping any issues mentioned above! A complete win-win scenario for all involved. And now we have the prompts perfected, we can create more and more images for them at a click of a button.
So, with the correct and ethical use of AI programs, there is no reason why it can’t become a key part of our creative lives. Any tool that opens up possibilities to be more creative, more innovative, and more efficient for our clients should be embraced with open arms.
And with NVIDIA, Adobe, Microsoft, Google, and many more jumping on board with AI, we’re set to get far better vector and text support, full motion AI generated video, and … wait for it … fully AI generated 3D models for use in Blender and the like!
A new creative future is upon us, and it begins with two letters: A.I. Now, if only they’d invent an AI platform to fill out my timesheets for me, then I’d be in heaven!