OpenAI’s ChatGPT 4.0 took a jump forward in the “Graphics by AI” request.
I entered this prompt:
“Create a graphic for a LAB about AI and Learning in the Workplace”
This is what kicked out of 4.0’s DALL. E 3 Graphic Generator - the graphic and it’s descriptor of what it designed:
I was quite stunned by the clarity, definition and responsiveness of the graphic. It added the “diverse group of professionals” and “a high-tech futuristic setting”. Plus, it added Learning below it’s AI logo and much more.
So, could I evolve the graphic. Someone shared that the age of the people were all “young and hip” - could we diverse on age. So, asked and here is a portion of the revision:
Interesting, but if you know Elliott, the idea of many people wearing suits and or ties does not fit. So, I asked for another revision and lose the ties and keep it more “hip”:
This is a leap forward and has way more functions to evole ve. But, the graphic quality and responsiveness was a leap forward. And, it will force use to evolve both our skills and the technology. As a learning executive, you may want the tech to:
Ability to click on a single person or icon and evolve that one with a prompt or substitute an actual image for your organization!
Get a full disclosure of the origin of the graphics. Were they “scraped” from existing images. Who owned the image? What are our rights to use the image?
Does Chat GPT learn about the requester. Can it look at Elliott Masie’s website and postings to extend and personalize the prompt?
Can it reflect the “Culture”, “Code of Conduct” or other policies of an organization?
Can the graphic sense the language of the viewer and perhaps translate the words and alter the images appropriately?
In addition, we need to look at the skill shifts for using this function:
Graphic Artists: How do they leverage AI to either create the first draft, revise an output image or reflect the graphic tone and style of the project or organization? How do they shift their fees when working with AI elements?
Learning Colleagues: Will we be able to generate i and learning support materials in graphic format? How do we add the art and science of learning to the shaping of these AI assisted workflow support tools?
Learners: Can we help learners generate AI images to support their learning process - reflecting their test resulsts or performances and highlighting focus with graphic emphasis.
People’s Images: Do organizations have the right to use the images of specific employees in graphics? Do they need permission? How long can they use it? Does the employee receive any fees? Watch for labor union negotiations on this issue!
Trust: What can we do in our design, publishing and teaching to help people trust the accuracy and “truth” of these images.
We are just at the beginning of exploring and experimenting with AI in general and now with the role of AI Image Generation. Much to learn and many issues/challenges to test and evolve in our organizations.
Keep experimenting with the various AI tools rapidly coming to the marketplace. These graphics were done with the paid version of OpenAI’s Chat GPT 4.0. There are also new systems from Google’s Gemini and watch for Apple’s announcements in the first quarter of 2024.
Yours in learning,
Elliott Masie
MASIE Learning Foundation
MASIE Learning Collaborative
www.masie.com
emasie@masie.com
PS: There are only 9 spaces open in our small AI & Learning LAB in Saratoga in January. Details and reservations: www.masie.com
I’ve been amazed at the quality of art as well, Elliott.
The big bummer I have, working with gifted, veteran, award-winning feature animation folks, is how their job opportunities may be adversely affected.
AI art I’ve noticed, though, tends to be too slick & sanitary. I’m hoping that missing “character” quality will always be needed — and valued / compensated — from human artists.
Elliott, as usual, you sent me down a rabbit hole. I lost hours to comparing art in GPT4/DALLE, Midjourney, older DALLE via Bing, Leonardo.ai, and a few others. I love playing with this and finding the weirdness in the AI images, as well as the fantastic.