AI the great leveler
Before we start, I feel like it is my duty to warn you about this essay. This is not a nuanced essay. The ideas presented here are very biased and are in favour of the development of Artificial Intelligence to its fullest length. Now that we have that little existential hurdle out of the way, I’d like to say I support artists and designers one hundred percent, I feel their pain and I promise to not be “holding salt to wound” while I type this (even though it may appear that way).
A light introduction
I am a designer myself, and for the last 7 years or more, I have looked at my screen almost everyday — save for a few days which are roughly less than 0.001% of the total — trying to create something nice for myself or for a client. Trying to understand my role as a graphic designer; then a logo designer; then a motion designer; then a product designer; then a 3D artist, tumbling back and forth these roles and the blurry lines in between them. I have followed passion deeply, and I have also followed the money that comes with the job. Both have been equally as exciting. Yet, 8 years after, I still struggle to meet function and the exact needs of clients and users in everyday design. I still struggle to create that great art concept and maybe bamboozle enough people to buy my NFT for anything more than 2 Ethereum.
Many artists and designers are like me. They love their craft and they’ve honed it for years. They’ve made money, they’ve done great stuff and have climbed the ladders of their profession one blank canvas at a time.
These creative people have spent the needed hours to get better at their craft and as a result, they’ve gotten better, and faster too. They spend only 14 hours to complete a 3D piece, that would usually take a newbie more than a week to complete to anywhere near that level of quality. You take a look at their work, and you see YEARS of screen hours in every detail. You see sweat. For these types of people I have the greatest respect.
Beautiful story, so what is the problem?
For many creative people who produce art and design AI seems to be the latest problem. Not client misunderstandings, not deadlines, not broken companies/employers; but a product of computers (one of their greatest tools ever), a computer baby. Artificial Intelligence.
You’re in for a small dose of history, so let’s get to it.
What is AI?
To define AI or Artificial Intelligence, we’ll need to find out what it’s not. Also, for simplicity, I’ll refer to Artificial Intelligence as AI throughout the remainder of this essay. AI is not human, since it doesn’t have a nice looking fleshy heart in it’s chest — if that counts for reason. AI is also not simply a computer, since it’s not garbage-in-garbage-out (GIGO). AI currently sits somewhere in the middle.
What this means is that AI, although not being able to be human with cells and blood, consciousness and all the fancy stuff, AI is able to fake it to a very large extent, while it rests on the computational power of traditional computing. This is a very dangerous combination.
Before I tell you why it’s a “dangerous” combination, I’ll give you a very textbook definition of AI below. This is from Britannica:
“Artificial intelligence (AI) is the ability of a computer or a robot controlled by a computer to do tasks that are usually done by humans because they require human intelligence and discernment.”
This means, when a computer gets smart enough to start thinking or acting intelligently — like, or close to a human’s intelligence — then it’s intelligence is artificial, since it is unlike what we’re used to and it is man-made.
So, it’s no different from a tool, like a Hoe or a Spanner. Maybe a Spanner that can talk back at you and show you what bolts to tighten on your Car. AI is a tool.
Oh, it’s a “small” problem:
How is AI development dangerous to the world?
About this time, 2022/2023, the smartest KNOWN AI in the world still operates at the Narrow Intelligence (ANI) level. Meaning AI clearly exceptional at one thing.
An AI that can play grandmaster level chess, Alpha Go; An AI that can serve as an assistant on your phones, Siri or Alexa; An AI that can drive and park itself, Tesla; An AI that can predict the next exciting short video you’d like to watch, Tik Tok; All of these AI exist today, and they are mostly very good at one thing and one thing only. Everyday, humans refine their code to make these technologies do much more. The more they do, the more they get closer to the next best stage of AI development, which is General Intelligence (AGI) — that’s us.
When they’re done with that stage, AI develops away from that level and reaches Singularity, a term popularised by Ray Kurzweil to represent the point where AI reaches Super Intelligence (ASI). Super Intelligence is a level humans cannot reach, and just like how humans cannot fly except aided by some kind of mechanism (a plane or rocket, etc.), humans cannot get to ASI level, except they’re carried or helped. Therein lies the danger of this wise tool. What’s the incentive of AI helping humans along to super intelligence when it gets there, when it can just out-think and out-maneuver our BASIC General Intelligence?
To learn more about this development, Tim Urban, who is the author of the WaitButWhy blog is the best of the bunch to understudy. Start your study from here.
How is AI development dangerous to creative people?
Remember we’re still dancing somewhere between Narrow (ANI) and General (AGI) Intelligence for AI? Well, even that level has become a threat today for some artists and designers.
The creative process takes time to come by and to create great work, you have to work very hard on it.
Time & hours = Great output
This doesn’t seem to hold true anymore, as AI of today is now able to out-compete many creative people. We’ll only talk about very few of the most trendy developments at the moment. Here are a few examples:
- DALL.E 2
- Stable Diffusion
- Midjourney
- ChatGPT
I’ll categorise Nos 1–3 under the Art/Image generation group, while I’ll put No 4 into the Research/Content/Syntax generation group. Let’s get to it.
Art/Image generation
New AI powered technologies like DALL.E 2, Stable Diffusion and Midjourney and so on, form the group of very competent — and that’s subjective — technologies that help people, I mean anybody, create any kind of image they can imagine. They do this buy giving the user an input field to type in a natural language prompt which generates an image.
These simple prompts may look like this:
“A cinematic photo of an astronaut nun” typed into DALL.E produces this. An image generated by Patrick Clair.
Asking Midjourney to create an experimental dashboard “for monitoring solar systems and nudging the cosmos intelligently in the direction of biological emergence” produces this and many more, generated by Mikael.
Including two other random works generated by Mikael.
A couple of layout and colour explorations for UI design by Davis Taylor Brown, using Midjourney AI.
AI can also be used to create really trendy client concepts for fashion. Here are some examples also created by Davis. He uses a combination of GPT 3, Midjourney, Stable Diffusion and possibly DALL.E.
And lastly, creative artists like Jon Finger, etc., use the power of AI to create movie scene concept and world building for AI based cinema.
These are just a few of the many thousands of artists creating with AI today. I look at these works and I see art, design and creativity in its purest form, and nothing else. Well, I lied; I also see speed. Speed matters.
Research/Content/Syntax generation
I’d like us to spend as little time as possible here, since we’ve already established the AI competence pretext. Let’s look at how natural language models like ChatGPT, Instruct GPT, GPT 3, etc., work.
This is what Open AI (they created ChatGPT) has to say about their creation:
“We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.”
It’s naturally the same or similar for most of the other tools mentioned. A user signs up, talks to the all knowing AI, and the AI responds with solutions, advise, feedback, in the most advanced way publicly known. I say “publicly known,” because I am not so sure what else has been created but is currently unreleased by tech giants like Google/Alphabet.
Designers, writers, artists or midwives could ask the ChatGPT for information about anything — any information, and it would provide an answer in a conversational way.
In using ChatGPT, I have found that I am very prone to playing mind games with it. But ChatGPT is much more. It can troubleshoot code and spit out clean code in any language (although, it’s not always right), it can write you a recipe, a poem, provide research information on a topic, suggest a title for your next YouTube video... Anything you want.
For example, here’s a scenario shared by the Twitter user @algoflows, it shows a short conversation with ChatGPT, asking it to reveal the process of hot-wiring a car, which the AI is reluctant to do at first.
Here’s another example where I try to get ChatGPT to introduce itself to you, the reader of this article. I also asked it to turn the answer into a rhyming poem.
It is important to note that an AI like ChatGPT is not perfect and can often make stupid mistakes, but at its current level, it’s still a sight to behold. You can learn more about how it works here.
Genuine concerns
Of course, these types of developments will always get people talking, so here are a few of the peak opinions, and some follow up thoughts:
1. Copyright issues
The first point of concern from creative people is that works created with these AI systems are sometimes too close to already existing works in the public domain. For example, you could tell any of the AI to “create a sci-fi scene of Delhi, in the style of Syd Mead” and it would create something seeming close enough to be believed to have been created by the now late Syd Mead himself. This appears to be very dangerous, since it can be done for any artist, in any style or a combination of two or more styles. You just need to be able to type and tweak the prompt enough to get what you want.
Eric a character artist and game developer, put it sarcastically, when he cries out in this Twitter thread. He thinks AI platforms are just outright stealing original work, without repercussions. A lot of designers in the thread seem to agree with this opinion, which brings me to the next concern.
2. Ingenuity
Other artists, like Dan Eder — a senior character artist in the USA — think AI art is neither polished enough nor is it qualified enough to be featured on the same portfolio boards as other artists who spent hours on hours creating real work from scratch.
Well, I have strong opinions here, which I will share at the end of this essay, but for the sake of our discussion, see the thread here for deeper — equally balanced — opinions.
3. Usefulness
You may say, yes, AI can create colourful UI design concepts, but how useful are these concepts in real world product design processes? Can AI create and maintain a design system? Can AI properly create a working userflow for your new tech product?
Many of these questions would currently come out with a solid no, but I daresay it’s only a matter of time before AI is able to figure these things out and get them to work.
Currently, it may not be useful enough to solve these difficult design problems, but it currently helps creators like Jon Finger in world building for movies and cinema. So, the defining factor here is time.
Famous last words
We’ve discussed all the talking points (or most of the pressing ones), but one thing stands true, AI is no different from a Hoe or a Screw driver. It is just a tool, and artists/designers need to get used to that and see it that way.
Once upon a time, designers couldn’t create any form of design, without a drawing board, some real life canvas, some real life ink, some rulers and very sharp knives. But these days, we can head to Adobe Photoshop, select the “fill with content-aware” and have Photoshop magically generate pixels from nowhere to help our photo-manipulation.
We can pick one of hundreds of brushes on Procreate or Fresco, throw pixels around that technically do not exist. We can look at hand references, feet references, torso references of real world people to make sure we’re getting our proportions right.
Product designers see a new interaction that works, and in a few months, every product in the market has that feature, fully tested and implemented. I can list Stories, I can list Reels.
We store reference everyday subconsciously (from nature and all around us) even without knowing it. Why is it so different when AI does the same thing?
What this means is that even the untrained common person, would be able to create art. You don’t need to have great motor skills, just have a great imagination and you’d be able to create with AI. What this means is that, even if you don’t have great imagination, AI will still augment, and you’ll create, better and faster.
Conclusion
As creative people, we literally create magic everyday, so why the jitter, when someone passes us a new and improved wand? AI isn’t going anywhere, get used to it.
AI artists are in fact real artists, and should be called just “artists.” The fact that they produce work faster than traditional artists or designers, without having to move a mouse as much STILL makes them artists. The move to put AI art in a box has failed even before it started. Again, AI is here to stay.
The best approach I have seen so far, is to take AI and fuse with it, create better with it, do things faster, get involved in its development and we’ll be all the better for it. Cheers.
Thank you for taking the time to read this article. Be sure to drop a few comments below.
