Creative Intelligence
Although there has been much teeth-gnashing and handwringing about artificial intelligence (AI) stealing people’s–and specifically creatives’–jobs, I would like to state my thesis upfront and clearly: Creatives, we cannot afford to bury our heads in the sand; we must participate in this movement.
AI is already here, and you cannot outrun it. AI companies have already used copyrighted work to train their models. Any creative endeavor posted to the internet (including what I’ve posted) could have been used to train a model or three.
Further, there is worry it will steal future work, as in, companies will no longer hire creative workers to make stuff. Maybe there is truth to that. Or, maybe we can learn to use AI to our advantage.
Intelligent Creations?
Companies offering AI are a dime a dozen. The most familiar among graphic designers and photographers may be Adobe Photoshop’s very useful Generative Fill tool.
Adobe, being the de-facto software used in creative spaces, offers Adobe Firefly, their own generative AI model. This was, obviously, trained on artists’ work, too. Photoshop’s Generative Fill runs on Firefly, as do its “Sky Replacement,” “Select Subject,” and “Content-Aware Fill” tools.
More broadly, many are familiar with the likes of Claude, Stable Diffusion, Midjourney, Perplexity, and Chat GPT. These can digest large amounts of data, generate charts, graphs, and images; and provide output in a specific format. These are all powerful and useful tools.
But that’s all they are: tools. As with fire, the wheel, and the internet: if you refuse to learn the tool, you choose to fall behind. The risk involved with ignoring AI is simple: you will be less productive, less competitive, and behind your peers.
Be an Intelligent Creator
Imagine if when the internet came out, designers refused to find out what it was, how it worked, and how to leverage it. There’s no displaying your portfolio on a website, no emailing contracts and getting them back–signed–faster than snail mail, and no easy way to pay subcontractors.
AI is a tool that is not going to go away, and creative workers need to gain exposure to these tools. AI has its benefits:
- Spending less time on monotonous tasks, more on creative ones; like cutting out the main subject from an image so you can move on to placing that cut-out into something more interesting
- Automating tedious, time-consuming tasks; like parsing a long document to identify key takeaways for sales sheets or slide decks
- Creating images where it is too dangerous, or impossible, for photographers to go; for example, a close-up, bird’s-eye-view of a waterspout approaching a densely populated coast
- Debug website code, like finding an extra space in your style sheet
To get started using AI, learn the basics of prompting. You don’t have to become a prompt engineer but understanding how the machine works through your prompt to create a useful output is essential.
State what you want it to do clearly, and break a larger, more complex prompt into smaller chunks. Do not be nice to the AI; tell it what to do, how you want the output formatted, and provide background information. Refine the output.
Experiment and iterate–that’s what creatives do!
AI as Uncreative Intelligence
AI can do a lot of what designers do, but let’s also be clear about what AI cannot do (as of now):
- It cannot typeset a magazine
- Generally, it sucks at generating pictures of legible words
- AI does not know how to apply color theory
- It still struggles with the “uncanny valley”
- AI has difficulty digesting long, complex instructions for large, complex projects
- And most importantly, an AI cannot make judgement calls
Without a doubt, AI will improve. The above list is not evergreen. Learn what AI can currently do, use it, and then use those future improvements to become more proficient.
AI as Creative Unintelligence
Learning to use AI sounds like “cheating,” and letting another company profit off the work you’ve already created is frustrating (people also thought using computers in graphic design was cheating, then they thought not hand-coding every website was cheating).
But unfairly profiting off your work is valid. Unfortunately, proving and finding recourse if a company used your work to train an AI sounds like a long, expensive legal battle with precedent against you.
You can, however, protect future work by poisoning the AI’s training data. You can try Nightshade to render visual work unusable if an AI company scrapes it for training data.
But keep in mind, this utility is not foolproof: researchers developed LightShed, which can identify if images have been poisoned and can remove said data poisoning.
LightShed successfully detected NightShade-protected images with 99.98% accuracy and effectively removed the embedded protections from those images.
utsa.edu
I repeat: you cannot outrun AI.
The Larger Language Model
Unfortunately, the argument can be made that AI further suppresses creatives’ professional valuation, contributions in the marketplace, and influence in culture. AI feeds into the long-held belief that artists’ work is less valuable because it is not measurable and does not have the rules of math or science.
Although everyone enjoys creative work–music, food, movies, books, clothing, interactive games–there still persists a culture around not fairly compensating for that work, as indicated by the following models of thinking:
- Artists must suffer to make their best work (e.g.: Edgar Allen Poe, Vincent Van Gogh)
- The “starving artist” stereotype
- Make it look easy, no matter how much it hurts
The creative arts make a society’s culture: what we listen to, eat, read, wear, and decorate our homes with. If we don’t learn to leverage these tools, we risk letting computers decide our culture–and that’s sounds pretty dystopian.