Exploring AI Through Hands-On Playful Learning

I welcome your comments on my latest experiment in leveraging generative AI tools in my creative process.

In this installment, I embraced DALL-E 3 to visualize concepts from my article, “Don’t Fall for the Hype or Hysteria about ChatGPT. Don’t be Complacent, Either. ” I used ChatGPT-4 for scripting, Eleven Labs for text-to-speech, D-ID for video generation, and iMovie to craft a highlight video. And, I used OpenAI's Creative Writing Coach GPT to tune this article. Total cost of professional assistance, software, and media = $36/month in subscriptions.

My aim isn't to proclaim the dawn of AI dominance over human creativity. Instead, I'm fascinated by how AI is pushing the envelope on tools for creative human users. To this simple end user, my few hours of tinkering with generally available tools yielded thrilling results. Imagine the possibilities of ever more advanced tools in the hands of seasoned professionals.

For example, I wanted an avatar for my video – something that was reminiscent of me but not a deepfake nor even photo-realistic. I started by asking ChatGPT to interpret my photo. From there, I refined the description and visual effects, and quickly crafted a DALL-E 3 prompt that generated this digital palette knife painting:

The other video images followed a similar quick iterative process. I started with a simple seed instruction to ChatGPT, which would then create a detailed prompt for DALL-E 3 to generate an image for my review. I would either give further high level instructions or edit the image prompt directly, and repeated until DALL-E 3 produced an appealing image.

For example, to illustrate the "don't be complacent" message, I began with a simple idea: a complacent, Monopoly-like character. Through iterative instructions and modifications, ChatGPT and DALL-E 3 helped me evolve this concept into a striking image that captured the theme of willful ignorance amidst technological progress.

To give you a peek behind one layer of the DALL-E 3 curtain, here's the final prompt developed by ChatGPT, in response to my instructions, that resulted in the above image.

Create a 16x9 image featuring a character similar to the 'Monopoly Man,' but in a pose where he is covering his eyes, as if consciously avoiding seeing what is happening behind him. In the background, depict a scene bustling with technological activity and advancements. The character's pose should suggest a deliberate choice to ignore or avoid acknowledging the rapid technological progress occurring around him. The contrast between his action of covering his eyes and the dynamic, high-tech environment in the background emphasizes a theme of willful ignorance or denial in the face of change and innovation.

To visualize our seven Laws of Zero , I prompted a series of images "showing experimentation, learning, and innovation" in each technology. I asked for them in different styles for visual contrast (and for fun).

This experiment was more than a technical exercise. It was a few fun steps in my personal journey towards understanding generative AI's potential, its limitations, and how rapidly it's evolving. It also equipped me to better assess the broader discussion on AI's role in our future.

So, while I won't be replacing professional art directors and illustrators anytime soon, the real takeaway is this: AI tools are rapidly elevating creative possibilities, and raising the competitive bar, for everyone.

My advice remains: be wary of the extremes in AI discourse. Don't get caught up in the hype, but also don't underestimate AI's potential to enhance our creative endeavors. In other words, Don’t Fall for the Hype or Hysteria about ChatGPT. Don’t be Complacent, Either ?(emphasis on the "Don't be Complacent").

And, by all means, invest some time in hands-on playful learning.

If you missed my first experiment in AI-assisted video narration, check it out here:

and read about it here:

https://www.dhirubhai.net/posts/chunkamui_welcome-your-comments-on-my-small-experiment-activity-7128815182834749441-izm8