How Much A.I. is Too Much A.I.?

I don’t blog enough to be topical or cover current trends, and I like to keep things light and funny, or positive and optimistic in general. In fact, at the end of 2022 when I actually tried to be topical and write about artificial intelligence, I did it from a pretty altruistic standpoint with the hope that working with the changing landscape that a.i. isn’t going anywhere is a better approach then railing against it. At my core, I still believe that, and I’ll put a pin in that for now, but today I’d like to revisit a.i. with a little more experience after having had the wonderful opportunity to visit the Adobe Max Conference in Los Angeles as an attendee.

At the Adobe Max Conference, October 2023 where I got to meet Aaron Draplin!

I say, “a little more experience” because not long after my initial blog post about a.i. in December, I followed that up in April of this year (I told you I don’t blog enough to be topical!) with a post detailing my impressions using ChatGPT and Dall–E which very quickly became dated after both platforms made serious advancements technologically. I think for most people at that time, a.i. was still a fun and funny little moment of pop culture that humorously added to the zeitgeist with things like Keaton Patti’s bananas Olive Garden commercials and Trump Rallies—all created by feeding hours of those particular brands original content into a.i. bot programs. Maybe the undercurrent had a slight worrisome tone of the inevitability of robot overlords, but it was still relatively light–hearted and quaint. Who could have ever imagined we’d be pining for the simpler times of 2022.

Of course, like most technological milestones, once something starts to get traction, it really takes off. Granted, a.i. has actually existed since the mid 20th century (starting with the Perceptron Mark I in 1957), but really started to generate public interest in the 80s with the goal of revolutionizing computer processing. It’s directly because of this that artificial intelligence has thrived recently—not so much because tech geniuses have learned more about a.i. themselves—but rather because computer storage, memory, and speed have increased beyond what most average consumers even need. And in 2023, Dall•E released its latest text–to–image model featuring significantly more nuance and detail, and ChatGPT became the fastest growing consumer software application which now offers it’s GPT–3.5 engine operating its services on a freemium model while ChatGPT Plus offers its GPT–4 engine to users for $20 USD a month. If that’s not impressive enough, as of this post, ChatGPT itself isn’t even a full year old yet having launched in November of 2022! In March 2023, Adobe released its generative a.i. tools to Photoshop in beta testing. Today, those a.i. tools are fully integrated into Photoshop and Firefly a.i. is in beta for Adobe Illustrator—creating editable and functional vector illustrations—as well as Adobe Premiere Pro which also offers beta tools for speech to text editing for videos. It is insanely easy to use and saves lots of time.

All of these images on Adobe Stock are generative a.i.

Let’s step back and address just a couple quick bullet points I made previously regarding a.i.’s learning technique. Now initially I made a mistake and thought the process was called stable diffusion, but that is in fact the name of a latent diffusion model developed by a company called Stability AI. Artificial intelligence uses machine learning to develop a deep knowledge of whatever subject it’s tasked to create. I’m cutting out a lot of context here for the sake of brevity, but imagine a robot who has the capacity to instantly read every single book on a particular subject so that it can then use that knowledge to compose its own creation based on that immediate education. It only has the information it has acquired, so it bases everything it can do on that information alone. I read hundreds of text books about various software programs when I was in school, but it almost all felt useless after I actually started working and realized experience was vastly more important. As a result, most of my initial professional work looks ridiculous, much like what a robot who only learned from reading and not actually doing.

Now imagine that same robot is instructed to paint a masterpiece, but in order to do that, it has to visually take in every painting currently on display in the Louvre. So it rushes around the museum and sees works from DaVinci, Géricault, Michelangelo, and Jacques-Louis David. Afterwards, you ask the robot to paint its own masterpiece, but upon completion you notice it hasn't really painted anything original, but rather cut and pasted elements like the Mona Lisa's smile or Roman columns from the Oath of the Hoaratii. It's specifically these issues that have a lot of creative people feeling pretty upset because the robot hasn't actually created anything, it's just stolen components from others. But then you also notice that because the a.i.'s creators have basically told this robot to go out and learn everything from the internet, you start getting into real troublesome areas because if you haven't noticed, there's some pretty horrible stuff online and it's not exactly hidden either. So now on top of being an art thief, the robot has also learned to be racist, sexist, and creepy. You know, like actual real life people.

Again, I'm really compacting a lot here to keep things from getting bogged down with technical jargon, but these are real concerns that have had companies like Adobe make serious public efforts to proactively promote responsible guidelines for generative a.i. learning and sharing, as well as protecting intellectual property and reflecting diversity in a positive way. For the most part, this corporate responsibility and good faith approach has been necessary, not just because of the reactions from creatives worldwide, but because of the accountability these types of organizations can be held to legally. Remember, theft of any kind is generally frowned upon.

So now we get to the Max conference where the undeniable star was Adobe's generative text–to–image a.i. program, Firefly. All over the conference floor, in its classrooms and displays, and promoted heavily at each keynote session were strikingly beautiful images all created by artificial intelligence. What Firefly offered was so prominent that it sometimes felt invasive. Is it cool and will it save designers from doing tedious things like masking, editing, and rough concepts that require super quick turnaround? Oh man, you bet your a$$ it will. But as one classroom speaker jokingly noted at the beginning of one session that it was the only conference event not promoting any new artificial intelligence tools, he was met with a roaring standing ovation.

Adobe competently showed that it's a.i. deep learning methods were trained solely on its own library of photographs, illustrations, images, and graphics (pretty much its entire stock library) and that its engineers were working tirelessly to integrate guardrails so that diversity and inclusion were represented equally and respectfully. So there you go! Problem solved. You can stop worrying now about everything. Robots are kind, love is love, intellectual property theft is a thing of the past, and the system works great.

Even if all that were true, there's still a hiccup or two. Now put your personal feelings about a.i. aside for just a bit and let's pull that pin out regarding my optimistic outlook from earlier. The cold hard fact is that a.i. isn't going anywhere, and just by looking at what was just a year ago and what is now today, it's pretty obvious that what a.i. can create is only going to get more impressive and it's going to be up to everyone to ensure it's guided properly, safely, and responsibly. Up until the Max conference, I could be heard saying, "Man, if this is where it's at now, imagine what it'll be like in ten years!" But that was out dated thinking when I first wrote about a.i. Especially when Adobe has expressed its expectation to see user generated a.i. images increase 5 times more in the next 3 years than it currently already has which is already over 15 billion!

So while ethically we have a lot of work to do, I think it's also fair to say that we're currently pretty early on in this saga and we're already experiencing some pretty heavy a.i. fatigue, and to that point, I'd like to redirect your attention to stock images. I love stock images, like I adore stock images and I've written about them before (and it's a funny article I'm really proud of too), but because I almost exclusively use Adobe Stock, and because Adobe Stock is the epicenter of Adobe a.i.'s learning process, it's kind of saturating the store with its own product and nothing else. Imagine going to your local grocery store, and all they sold was their brand of corn flakes. No produce, no deli, no butcher; just aisle after aisle of varying sized boxes of store brand corn flakes and nothing else. Because Adobe's generative a.i. has come along so far and has gotten so good, it offers generative a.i. images as stock image options. Originally you'd see one ore two pop up, then it became the majority of what was offered. Now depending on what you're looking for, a.i. generative images can be all that's available.

Real quickly if you haven't read my post on stock image sites; they provide a designer access to photos, graphics, templates, or illustrations the designer wouldn't otherwise have time to create themselves. Creating an ad for a new coffee chain? I can search for something like, "Friends enjoying coffee together in a cafe" on a stock image site just like I might look for something in a search engine and I will get various results that will hopefully match the look and vibe I'm going for.

So just to clarify, I don't have an issue with a.i. generated images. The quality is really good and getting better. There will be times when an a.i. generated image is much better than anything else the stock image site is offering, but it's frustrating how much of it there is. When searching for "Friends enjoying coffee together in a cafe", the language in the search itself is referring to a very human experience. The generative a.i. image is good, but it's not perfect. Plus I think there's a subconscious bias that I personally have that (at least in this instance) using something that was not created by humans but that is supposed to be representative of a human experience does not feel like a genuine, intentional choice to encourage others to buy into the design I'm making.

This image was generated by a.i. Can you tell?

I realize there's so much to unpack there. I'm using a computer to design this resource, is it really that bad that I'm requesting a computer create an additional element to that? Will future designers be less likely to have such a bias if they grow up understanding how ubiquitous this technology is? How "human" does an image have to be to properly reflect a human experience when we're all already familiar with shorthand cues like seeing people in such settings—photographed, illustrated, or otherwise—that creating that established connection is even necessary?

For their part, Adobe has a very clear and up front policy regarding their ethical standards and practices regarding posting, hosting, sharing, and creating generative a.i. images. Now people will point out a lot of that responsibility falls heavily on the users respecting that system, and this is what I personally believe is at the heart of all of this and creates the endless loop of debate surrounding this. Policing people from playing with this technology may be noble, but it's also antithetical to creativity in general. I'll repeat it again:

Artificial intelligence is not going anywhere and will continue to develop, but a.i. is a mirror reflecting back on the people who use it, interact with it, and engage others with it. Just like the world we live in, what we put in will be synonymous with what it gives back.

A.I. images generated from Dall E 2 and Firefly using the prompt, “Renaissance painting of a black cat in a pink tu tu”

But what do you think? Is a.i. the cool new future that will bring the world peace and prosperity? Is this the beginning of the end? Will a.i. steal jobs from creative people as well as blue and white collar workers? Or am I just fixated on the loving memory of our cat Destiny and want a.i. to realize how special she was like my wife and I do?

Our actual (late) cat Destiny killing it better than any artist—a.i. generated or otherwise—ever could. Rest in Peace, sweet girl.

Thanks so much for stopping by, and I really would love to hear your thoughts on this. Please follow me on Instagram and let me know! Also like last time, here are more sources as well as some other great articles about a.i.:

Tedious "To Do" List Lunacy

I was so bad at doing chores and keeping up with my homework as a kid that my parents got me a chore checkoff pad. If you’re old enough to do so, think of a yellow legal pad, only each tear–off page is just the same pre–printed copy of a ludicrous amount of “chores” with check off boxes categorized into different lists, and a bunch of side bar like areas with blank spaces to write in notes and even your own thoughts or special agendas. It was the early to mid 90s and this was a novelty that could only exist before computers and smartphones. Just search “To Do List” in the App Store and you’ll get dozens if not hundreds of Task Manager related hits with more features and personalization resources than a cheap scratch pad from Kmart could have ever offered back in the day.

As a Gen Xer, I’m still genetically predisposed to relying on old fashioned, handwritten or printed media as well as on modern technology in this weird dual reality, and it was that frustratingly detailed chore checkoff pad that serves as today’s inspiration. I keep multiple lists—both physical and digital—that essentially can be broken down into three categories: Things I Need To Do (long term), Things I Want To Do (goals), and Daily To Do lists (aka, “This $#!% Needs To Get Done Now!”).

I could write a lot more extensively on the benefits of these lists, how it’s satisfying to physically drag a pen over an item and cross it out resulting in that tasks completion, or how the whole process is really just a way to compartmentalize life without freaking the hell out, but instead I’m going to write about the absurdity of the second category mentioned, Things I Want To Do and its more specific subcategory, Things To Draw. Now I actually have several of these lists—some more unconventional than others—but they all serve a very specific purpose which can basically be explained as “Staying Creative Insurance.” All creative types know and struggle with creative blocks, and what’s more, they know that creative blocks will always, always be a certainty. So to preempt those blocks, creative people will have “lists” of some kind to assist breaking up those stale moments. This act may be subconscious as I’m sure any number of artists, writers, or Fashion Taxidermists reading this are thinking, “Hold on, I don’t keep lists like that!” But I can assure you that they/you do. A “creative to do list” can take the form of bookmarked content on social media, specially labeled photo albums, Pinterest Boards, bookmarks on your browser, screen captures, a specific shelf in a personal library, scrapbooks, following other artists for daily inspiration, or praying to the Norse God Bragi for guidance on a plausible excuse to post all the hotdog related art you’ve ever done (so far). A “To Do list” or “inspirational collection” then really becomes a matter of semantics.

The other tiny tidbit you should know about these lists—in whatever form they take—are how unique and personal they are to their creator. They will almost always only be decipherable to their author (mostly, but I’ll expand on that more in a bit). We all have our own little idiosyncrasies with everything we do, but when it comes to something more personal like a list of things to do created expressly by and for the only person in the universe designated to see them, those itemized collections can be downright hysterical even if their context is relatively clear (which often it is not). For example, to this day I am haunted by a very old to do list of things to draw with one item simply listed as “Date.” The confidence I must have had when I wrote that is staggering! Was it in reference to an outing I had with a girl? A specific holiday? Did I think there was something to add to Salla’s quote in the monkey death scene in Raiders of the Lost Ark? What’s even wilder is I actually remember the physical act of writing that down, but not the context as to why. If that wasn’t infuriating enough, I kept that drawing prompt idea on said list for years in fear that if I did remove it, I may lose whatever inspiration it ever possessed! It was important enough to write down for some reason.

Case in point, I’m at a place in my life now where even if I suddenly had unlimited time and motivation, all my to do lists would perpetually exist in the same state as my Netflix queue; I’m not going to get to most of it, but I’ll never clear either out as both stand as monuments to my interests if not as one to my goals.

For the purposes of this post however, I’ll be sticking to two “lists” of mine in relation to wanting to draw something very specific. The first is what I’ll call my “Definitely Draw This!” List which is literally a collection of pictures, and screenshots I collect and add to regularly. Incidentally, when I think of my actual “Things To Draw” list, it’s this collection of images I consider as the main one. It’s on a small thumb drive in a folder simply labeled •To Do. The bullet point ensures that no matter what else I save to that thumb drive, that particular folder always remains at the top. A quick browse through it shows mostly images of dynamic poses, facial expressions, architecture, and design styles I want to illustrate. Many of them I already have too, but like my Netflix queue, I don’t remove them after I’ve accomplished what I wanted to. In most cases it’s because those images continue to serve the inspirational purpose as well.

But like my “Date” prompt I mentioned earlier, there are a few pieces that make me wonder why on earth they’re in this folder. Some may take a minute to get (Did I screen cap this book title because I wanted to read it? No, it’s the typography… yeah, that’s it!) Honing in on features, hairstyles, fashion choices, or even specific style choices usually helps, but occasionally I’m at a complete loss.

My second list is more of the scraping–the–bottom–of–the–barrel variety. I love, love, love to write; but art is visual and it’s easier to save a picture with a clear mental note (e.g. “mimic this line art style”) than an overly detailed and usually tedious written direction. If I have an idea that I don’t yet have a visual idea for, I write it down as thorough as I feel I need to in my Notes app. In some cases, this can be relatively short as the prompt in and of itself carries those personal idiosyncrasies I mentioned earlier. For the most part, these have invisible, secondary ideas attached to them like a password or riddle that I and only I would understand. It doesn’t always work as in the “Date” scenario, but usually it does. For example, the prompt “Updated Hamilton” might make most people think I want to do my own take on Lin–Manuel Miranda’s musical, but in actuality it’s an idea to resurrect my own stab at comic writing where one of the antagonists from my strip Lazy Comics is named Hamilton. More detailed prompts like, “Italian/Greek girl, prominent bigger nose, hair in bun on head almost like Egon Spangler” give a pretty clear description of who to draw but offers no context as to why. I know however that this specific caricature was supposed to be part of a series of colorful real life people I observed down the shore when I went to the beach. I certainly won’t take pictures of complete strangers without their permission, and I also know (that in this particular case) I’ve already created a specific style and look, so descriptive written features will allow me to more boldly caricature someone without directly capturing their specific likeness. As for “Fist for a nose”? That’s a “Date” reference again.

What type of lists, prompts, pictures, or collections get you to stave off creative blocks, bring you inspiration, or help remind you to cancel that YouTube Premium free trial? Let me know here or @ me on Twitter and Instagram!

A.I. Blog Post Fun Time

Last year I wrote a blog post about A.I. that dealt with how stable diffusion lead to claims of art theft. I’m really proud of that article and you should totally read it here. In short, I took a different stance than most of my peers and tried to approach the subject from a slightly less biased angle and looked at the idea of trying to encourage other artists to embrace A.I. or at the very least, try and find a way to live along side it.

Today I thought I’d put my money where my mouth is and make a post almost completely made by artificial intelligence. The images are all created by Dall–E 2 and the post is written by the controversial Chat GPT. Now I’ve messed around with Dall–E 2 before but this is my first time using Chat GPT. I figured the main reason I started this blog was to reflect how I have grown as a creative and as a father/uncle and how kids help foster creativity in adults. So it seemed apropos that I should ask Chat GPT to tackle that very same thing and then pepper in images from Dall–E 2 with the prompts like: “creatively stifled adult finding inspiration by drawing simple pictures with a child” and “a children's drawing of a busy, frustrated adult.” Here’s how it went:

Chat GPT: Hello! How can I help you?

Dave: Please write a blog post for my website https://www.davehulteendesign.com/blog

Chat GPT: Sure, I'll be happy to write a blog post for your website. What topic would you like me to write about?

Dave: Finding creative inspiration by drawing with children

Chat GPT: Great topic! Here's the blog post:

Finding Creative Inspiration by Drawing with Children

As adults, we often get caught up in the daily grind and struggle to find creative inspiration. Whether you’re a professional artist, graphic designer, or just someone who enjoys doodling in their free time, it’s important to find ways to stay inspired and motivated. That’s why drawing with children can be such a magical and fulfilling experience.

Children have a unique perspective on the world, and their imagination is boundless. They see things differently than we do, and their creativity is not yet limited by experience or practicality. When you draw with a child, you’re invited into their world, and you get to see things from their point of view. This can spark new ideas and perspectives, and help you see your craft in a whole new light.

But drawing with children is not just about finding inspiration. It’s also about connecting with them on a deeper level. Drawing is a form of communication, and when you draw with a child, you’re sharing a moment of creativity and expression. This can build trust and foster a deeper relationship, as you’re creating something together that is unique to your bond.

So, how do you go about drawing with children? Here are a few tips to get you started:

  1. Choose a simple subject: Pick a subject that’s easy to draw, such as an animal, a flower or a tree. Avoid complicated shapes or structures, as it may be too overwhelming for a child.

  2. Let the child take the lead: Allow the child to choose the subject, the materials, and to create the initial sketch. This will empower them and make it more fun for them.

  3. Encourage experimentation: Children love to experiment, so encourage their ideas and let them explore different colors and techniques. Be open to their suggestions and don’t be afraid to try new things yourself.

  4. Make it fun: Keep the session light and fun, without any pressure to create a masterpiece. Think of it as a playful exercise in creativity, not a high-stakes art project.

Drawing with children can be a truly rewarding experience. It allows you to tap into your own inner child, while connecting with a younger generation. It’s a chance to create something beautiful together and to inspire each other to keep on creating.

So next time you’re feeling stuck or in need of a fresh perspective, grab a piece of paper and some crayons, and draw with a child. You may be surprised at what you’ll learn, and the inspiration it may spark.

Okay, Dave here again, and aside from being a little bland? I gotta’ say I’m impressed! Especially the 4th tip to “keep things light and fun” and not think of it as “a high–stakes art project.” The Dall–E 2 images are also not too shabby! I especially love this one; look at that kid’s face! He’s so angry!

Dall–E 2 generated image

The reason I was really curious about trying this out was A: It seemed like an easy lay–up that would require minimal work on my end (meh, not so much), and B: I’ve been seeing lots of ads for A.I. writing blog posts. I wanted to dig just a little deeper and find out if this is wide–spread and the answer seems to be, Maybe?

According to an August 2022 article on Story Chief Insights, Dayana Mayfield writes that search engines like Google tend to be negatively biased towards A.I. written posts because they are very derivative, don’t represent a brands specific voice, and can produce false information. Now most bloggers who do use A.I. use a service called Jasper which allows for more control and uses a more intuitive A.I. Jasper also boasts that major companies like IBM, iHeart Media, and Harper Collins use their services, which lead me to believe this may be more ubiquitous than I assumed, but finding exact figures is understandably hard to determine. That being said, most media outlets feel A.I. is better used by assisting writing rather than replacing—a consensus that seems to be echoed by graphic designers the world over as well.

I personally write because it’s therapeutic rather than being content driven; my three month hiatus is proof–positive of that. However, as I’ve said before, I do see the benefits of A.I. and try to stay open–minded while being responsible. I mean look at these A.I. generated pictures of a black cat I had Dall–E 2 create! Are they good? Meh. Are they entertaining? Absolutely.

Please follow me on Instagram and Twitter and thanks for stopping by! Please let me know your thoughts on A.I. and how it contributed to this article.