On AI

For many years, the letters “AI” and the phrase “dot-a-i” have referred to Adobe’s widely used application named Adobe Illustrator, and to the file name extension for the files we create using that software. But from now on, we’ll need to clarify which one we’re talking about: Adobe Illustrator or Artificial Intelligence.

This is a discussion of AI (as in Artificial Intelligence) centered on generative AI for imagery rather than for text or audio generation, or any of the MANY other rapidly expanding uses to which AI is (and going to be) put (list too massive to include here). And this discussion is only one more opinion among millions.

Given the insane pace of this technology, everything I say here will be out of date … oh, about 10 minutes ago. So, in considering this topic, I did have to ask myself “Why bother?” It’s difficult to have perspective of any kind when you’re in the midst of an “explosion.” As of May 17th, 2024, there were over 24,000 AI startups in the US alone.

Generative AI is at its core a copying tool. It creates output formed of mimicking and remixing. It also begs a question: What Is the Fate of All Tools? — One possible answer: No human invention exists that has not been bent to nefarious use.

The impact of AI, as we all know, is profound: sociological, psychological, legal, technological, economic, political, ideological, inspirational, existential. Destined to be bigger and faster than the Internet and social media, it collides with and influences those as well. It is dizzying, double-edged, controversial, game-changing, challenging, and certainly disruptive. Is resistance futile? Ask your friendly neighborhood futurist.

AI will write your love letters (step aside, Cyrano), your contracts, your business and marketing plans and pitches, your blog posts, books, and movie scripts … I’ve already seen a system that will write your e-book soup to nuts including text, layouts, and illustrations. I have zero doubt that one day — combined with robotics — AI will cut your hair.

Yet in the face of all these negatives, we know that the flipside of scary is exciting. We are indeed in “Interesting Times.”

Early users are still noticing that AI-generated images of human beings do NOT render hands, feet, or legs (and sometimes faces) very well at all, often producing monsters with strange quantities of fingers and toes, or deeply distorted features. I prompted (using an AI which shall remain nameless) for a scene showing a log cabin by a lake, with a pier and a float plane next to it. Got cabins with wings jutting out of the roof and everything from an indecipherable quasi-drone to a fighter jet. A mishmash that only AI could produce. I’m reminded of the old saying that to err is human, but to truly foul things up requires a computer. Then again, the results I was shown could be inspiration for a 21st-century Magritte or Dali (to clarify here: not DALL-E, but Dali of the very-organic Salvador persuasion **).

It is certain that working designers and artists — as well as students — are passing off AI-generated work as their own in an effective quantum leap of poaching that goes far beyond mere plagiarism. There’s that inevitable nefarious use. And with that use, how will the hapless wetware tell the difference?

I’m struck by the recursiveness of it all. We’re way beyond what I call the “talk-to-the-hand effect” that’s already endemic to social media: where ghost-written posts are out there getting read by no one. With AI, previously created content is mined for current AI-generated results. Current and future AI-generated results will in turn be mined for further new results until it’s all one huge mass of whatever-it-is built on a perpetually remixed copying of copies.

The mind boggles.

For many, AI represents a trivialization of the creative process that blurs both creative and ethical boundaries. I’m talking about the truly creative process that produces art by humans, for humans. The early players in this arena have exploitatively and unethically scraped the Internet for images both in the public domain AND copyrighted, without permission or payment. The courts and legislation are having great difficulty catching up, let alone “getting ahead of” this technology. Licensing, certification, fair use criteria, any regulation of generative AI is far behind the curve. New/future AI models will hopefully be trained from scratch, untainted by the work of bad-faith rule-breakers. Still, I need to ask: will an AI model generate the text for the class action lawsuit that’s filed on behalf of human creatives?

Is it too late for us to hope to cope fairly and effectively with this Pandora’s Box? The proverbial cat, as Kramer once said, “Mmrrrooowwwrr, is out of the bag.” ***

Job displacement is a certainty. For designers: Will AI eat your lunch? And your breakfast? And then your dinner from aperitif to dessert? It’s certainly snacking aggressively on our work “as we speak,” not waiting for an invitation to the feast. The designer’s response to this ultimate challenge of our agility as creators and businesspeople demands —

A need to think more as art directors and managers of tools.
The acquisition and mastery of new skillsets not only in composing the prompts (text or spoken, or whatever the interface evolves into), but also in the evaluation and decisions we’ll make about the use of AI-generated input to our own processes.
More of what remains unknown in the bright but murky future.

For us humans, our superpower lies in understanding the still-human aspects of what’s served up to us by our world, including by AI. Humans still have the edge on “vision” (both biological and metaphorical), on critical thinking, on awareness (computers haven’t “nicked” that … yet), and on sensitivity to the human audience (assuming that the work designers do is still consumed by other humans!). We corner the market on insight, on understanding the elements of viewer responses, needs, and requirements — all with nuance and originality. Machines know nothing of taste, interpersonal skills, the emotional mission, the organic aspects of the message, or the intangible goals of the designs we produce. Human minds “connect the dots” with synapses, not syntax.

As a tool, AI can produce multiple iterations and versions of imagery useful for fast ideation, a digital “mood board” of sorts, useful as input to human creativity — subject to human understanding. AI results can feed divergent thinking and unexpected (read: “accidental”) inspiration — including a log cabin sprouting wings. Proponents of generative AI focus on its “assistive” aspects. They say that AI will relieve designers of “all those laborious, repetitive tasks.” Rest assured; my designs do not involve “laborious, repetitive tasks.” My hope is that humans will not lean on, but rather leverage AI to enhance, augment, assist, but not to supplant us.

The rise of generative AI brings with it prospects for an entirely new line of work. What titles will we give to these emerging jobs? AI “Technician,” “Wrangler,” “Master,” “Jockey,” “Guru,” “Practitioner,” “Specialist,” “Virtuoso”? Regardless, those who offer their services with the special skillset needed to write, revise, and refine the needed prompts will bring with them a whole new bag of tricks of the trade, terminology, techniques, and best practices as power users.

For now, count me not as a Luddite, but as an engaged observer and occasional user. Both AI and humans have a steep learning (training) curve ahead of us.

Maybe AI wrote this entire article.

References and Links [Prepare to Be Overwhelmed]

“How Many AI Companies Are There? (2024)”

“How Many Companies Use AI? (New Data)”

A tiny glimpse into the pros and cons

References and Links [Prepare to Be Overwhelmed]

See also