I know, it’s incredibly controversial. My artist friends have been up in arms for a few years now over AI. It’s robbing them of their livelihoods, and they are right to worry so.
Confession: I stopped drawing two years ago because of it. I’m truly a frustrated artist. I knew a long time ago that I would never be as good as I wanted to be. I saw the writing on the wall that it would quickly surpass anything I could do. And I was right.
I’ve seen the arguments. It takes money out of the hands of artists. There’s no skill to it. If you use it, your ability to do the things that it does will erode. Well, that’s okay. What I’ve discovered so far puts the lie to that. A month of so back, my gaming buddies and I were chatting about its use. I was about to start running a new Star Trek Adventures game, we were starting all new characters. Just for fun, I thought, let’s see what 1966 Charlton Heston would look like in a Star Trek Uniform. I have to tell you, it didn’t do a bad job. Now, was this something I was likely to commission from an artist friend. No. No way ever.
Now, I quickly discovered that ChatGPT does not like to do the correct insignia on its Star Trek renderings, so I used my skills with Photoshop to Clone tool away the patch on his shirt and replace it with a photograph of the actual command insignia. Since I didn’t ask for a background I was free to place him wherever I wanted him.
I put him in front of a control console on my starship, the USS Nautilus, and interior set that I paid a whole lot for. Now, I don’t want to hang this photo up. I just needed it for NPC reference in a game. And if ChatGPT can give it to me and I can complete it in 15 minutes, then sign me up.
I used it to create backgrounds for my two Bill Krause-designed Challenger-class starships. This is also something I would never pay anyone to do. And I paid $400 for my two digital models, thank you very much.
So, yeah, while I’m using it for art, I am still paying artists for their work, but where else am I going to find a thousand backgrounds to put my ships on to my specifications?
The real kicker though, came when I was in a professional development meeting on June 6. We were learning how to make our home pages in Canvas and I asked the tech person how to word wrap around my photo image in my introduction. She quickly said, “Canvas doesn’t do that. But we can have ChatGPT write the HTML code for doing it and use its HTML editor to make that happen. And we did just that. Everyone, I have no intention of learning how to code HTML. This was a great tool to use and it never even occurred to me to use it for anything but playing around with photos and Star Trek stuff.
Then when someone suggested writing test questions with it, I had to admit, I didn’t know it could do that. But oh, yes. Here’s what I gave it:
Write a multiple choice question using Indiana Academic Standard 7.RP.2 with a Star Trek Original Series theme.
ChatGPT said:
Sure! Here’s a multiple choice question aligned with Indiana Academic Standard 7.RP.2 (Recognize and represent proportional relationships between quantities), with a Star Trek: The Original Series theme:
Indiana Academic Standard 7.RP.2 – Multiple Choice Question
Theme: Star Trek – The Original Series
While exploring a distant planet, Captain Kirk notes that the amount of dilithium crystals mined is proportional to the time spent mining. In one mission, 15 kilograms of dilithium were mined in 3 hours.
If the crew continues mining at the same rate, how many kilograms of dilithium can they expect to mine in 8 hours?
A) 35 kg
B) 40 kg
C) 45 kg
D) 50 kg
Correct Answer: B) 40 kg
(Explanation: 15 kg / 3 hr = 5 kg/hr. 5 kg/hr × 8 hr = 40 kg)
Now, I don’t know about you, but my day is made. BUT–Is Captain Kirk really the right one to note the speed at which dilithium crystals are being mined? No. That’s where my brain kicks in. No, it would be Scotty. So, I will make the adustment.
While exploring a distant planet, Chief Engineer Scott notes that the amount of dilithium crystals mined is proportional to the time spent mining. In one mission, 15 kilograms of dilithium were mined in 3 hours.
If the crew continues mining at the same rate, how many kilograms of dilithium can they expect to mine in 8 hours?
A) 35 kg
B) 40 kg
C) 45 kg
D) 50 kg
Now I have a decent question, and it took a fraction of the time it would have taken me to come up with something similar.
So, will my art skills atrophe? I don’t honestly care. I’m not doing this for money anymore. I’m more interested in making my workflow more efficient. My other skills, like using Photoshop, are getting a real workout. Especially when I want to see what Alan Rickman would look like as a Vulcan doctor. I fixed the insignia, the eyebrows, and I placed him in another set that I paid to have designed.
I know, all of my artist friends are dead set against it. I don’t blame any of you. But how AI is going to help me is immeasurable. I don’t care about six-fingered hands. What I do care about are things like I did yesterday. “Write a diagnostic multiple choice quiz with four questions from each of the five domains of the 6th grade Indiana math standards.”
Done.
I looked it over and it was solid work. If there were any poorly worded questions or questions that didn’t measure what I wanted to know, I could tweak them, but I didn’t need to.
“Write the quiz in Spanish.”
Done.
Then ChatGPT asked me if I wanted a printable PDF or if I wanted it in digital format. I chose digital format. Then I asked to make it for Canvas. It not only put the quizzes into digital format, but also gave me step-by-step instructions on how to upload it directly into my Canvas course.
So, in less than ten minutes, I had quizzes in Engish and Spanish uploaded into Canvas and ready to be assigned. Care to guess how long all of that would have taken me to do manually? Especially the translation? Hours.
Now, the debate can begin. Is it fair to rely on AI to do our jobs? Of course not. But I am a trained professional wth 38 years of classroom experience. I know good questions from bad ones. I also know if something is built to tell me what I need to know or not. I don’t rely on AI to do everything. I am going to use it to do the grunt work that I could pay an assistant to do.
I used it recently to write Ebay descriptions for comics and action figures that I put up to sell. I had to be very descriptive in my prompts. And I am going to say this, and you can argue with me all day long. It is an art form (and maybe a science) to construct a prompt to make the AI do what you want it to do. I have had many false starts because I wasn’t precise enough in my language when creating prompts. I’ve gotten a lot better at it now that I understand what it does and does not do well.
Does anyone remember the episode of Star Trek The Next Generation where Geordi gets tired of Data beating all the Sherlock Holmes mysteries on the holodeck and instructs the computer to create a unique Sherlock Holmes mystery with an adversary who is capable of defeating Data? This is how we got the Moriarty villain. You have to choose your words carefully.
But that’s not the only time that we’ve seen how AI works in our popular culture. When Tony Stark interacts wtih Jarvis or Friday in all the Iron Man and Avengers movies, that’s exactly what he’s doing. I recently had that same experience.
I started working on Heroes of STEAM again recently and uploaded the rules into ChatGPT. It gave me feedback on the gamel and how it was appropriate for the audience. I had the AI create two characters using my ruleset, and then had it simulated a battle between the two characters. It gave me a readout of the mock battle, all the moves, the damage done and the damage mitigated, and told me who the winner was. It even asked me if I wanted it to give the characters names. They weren’t bad, to be fair. I did note that it wasn’t having the characters move on the map at all. They just stod next to each other and slugged it out. I then added the strategy of moving to a distance where the character’s best attack would work. That generated an entirely different result. Then I had it create five characters and did a free-for-all fight. Last one standing wins. I ran that simulation ten times and found that the character whose main trait was mobility, the ability to dodge and move, was winning nearly every time. I asked ChatGPT how I could modify the stat to even things out a bit, and it gave me a potential solution. I made that change and reran the simulation ten times again. It changed, but now favored a different attribute.
I did this back-and-forth for about three hours, running hundreds of simulated battles using my rules. And now Heroes of STEAM is a much better product. The time and effort that it would have required to playtest hundreds of games was reduced to about 1% of what it would have been.
People can argue against AI all day long, but I’m going to tell you this; I’m not going back. It’s a whole new world now. It’s scary, but there’s no way this tool is going back in the box.
I remember when the first computer gradebook was released and my mentor teacher urged me to use it. It took me far longer to input the grades than it took me to calculate them mentally from my hand-written gradebook. I told him then that when the technology saves me time, I will start using it.
That day has arrived.
