Exploring the Potential of Dall-E 2 in Educational Settings
Written on
Recently, I was fortunate enough to gain early access to Dall-E 2, an innovative AI tool highlighted on the Open AI website as capable of producing realistic images and art based on natural language descriptions.
Having watched enough episodes of Black Mirror, I am well aware of potential pitfalls, including significant ethical considerations. Yet, my contemporary educator's mindset, accustomed to the casual acceptance of intriguing technologies, sparked with imaginative and abstract prompts.
A Start with the Official Prompt Guide
In the UK, the longest-running radio program is “Just a Minute,” where participants must speak for sixty seconds on any topic without pausing, repeating, or deviating. I approached Dall-E with a similar mindset, letting my thoughts flow freely:
> A teacher sitting on an infectious disease wearing a face mask surrounded by inspectors
Creating this initial bizarre and amusing image revealed two key insights: first, significant effort would be necessary to achieve satisfactory results; second, generating prompts in such an unstructured manner might inadvertently expose my subconscious thoughts. It resembles a reverse Rorschach test. Remember that? The inkblot test you’re thinking about now? Allow me to demonstrate with Dall-E:
> Frustrated teacher Rorschach inkblot
With the traditional Rorschach test, you interpret abstract inkblots, revealing insights into your personality and emotional state that are often obscured. Dall-E, conversely, prompts you to unveil inner feelings with a simple word choice, resulting in an abstract image that highlights emotions you may not have previously acknowledged.
After experimenting with a few of my 50 free Dall-E credits, I turned to the official prompt guide, which instructs users on optimizing the tool for manipulating aesthetics, vibes, emotional language, photography, film styles, illustration styles, and more.
Crafting Noun Phrases with Flair
I resumed my efforts by generating straightforward noun phrases centered on “teacher” and adding an artistic style. What began as a fun exercise produced unexpected results:
> Stressed teacher in a field linocut
The emotions captured in Dall-E’s four variations felt strikingly familiar. Is that how I truly perceive myself?
Why did I initially choose the adjective “stressed”? And why a field? It’s late August in the UK, meaning a mere week remains before school resumes. Shouldn’t I be preparing for the upcoming term? My “teacher” identity tends to surface unexpectedly, sometimes overshadowing moments of relaxation in nature.
What if I altered the style to reflect an artist I admire?
> Stressed teacher in the style of Keith Haring
Nope. This still feels too much like me. Let’s try a different genre.
> An exhausted teacher, by Katsushika Hokusai
Well, this doesn’t resemble me at all. Yet, I empathize with his experience! Perhaps shifting the subject will offer a different perspective:
> A politician in a classroom in the style of antique grunge
Not me either, but representative of a character that often appears during lessons. The antique grunge style evokes a historical context of psychiatry and psychoanalysis, embodying the inspector archetype.
> A nervous teacher, by Kandinsky
The Concept of Artificial Appropriation
Incorporating “by artist” in a prompt invites Dall-E to emulate the style of a known artist, often yielding whimsical results, especially when contrasting timelines:
> Monkey in outer space riding on top of a spaceship, by Katsushika Hokusai
> A student in a Tesla in front of a school building, by Paul Klee
> Sonic the Hedgehog by Van Gogh
While results can be hit-or-miss, my prompts featuring Shaun Tan, the esteemed author and illustrator of The Arrival and Tales from Outer Suburbia, produced surprisingly authentic images. With a little imagination, they could easily pass for his work.
> A boy on a bicycle by Shaun Tan
Tan’s illustrations are often surreal and convey social commentary that extends beyond what is typical in children’s literature. While Kandinsky, Klee, and Hokusai can’t voice their opinions, Tan can. I wonder how he feels about an algorithm mimicking his style? Does it irritate him, amuse him, or inspire him? His illustrations already possess a randomness reminiscent of machine-generated outputs. Is the algorithm a threat to his artistry?
Here’s the cover illustration from his book Cicada, an authentic creation born from a human mind and hand:
Alright, top that, Dall-E 2!
> A cicada wearing a grey suit and tie holding paper on a grey background, by Shaun Tan
I doubt Shaun Tan has anything to worry about.
However, what about lesser-known artists? I'm a fan of geometric and line art, particularly the Texas-based artist Tyler Foust.
Could Dall-E’s algorithm overshadow his style, hindering his artistic journey?
> United Kingdom outline, line art
For now, I think Tyler is safe!
A Classroom Resource
Last year, while experimenting with Adobe Premiere, I combined royalty-free images and music to create a video of Robert Frost’s renowned poem, “The Road Not Taken.” I wonder if I could have instead input each line into Dall-E. Let’s explore:
> Two roads diverging in a yellow wood, warm lighting, wide shot, outdoors, digital painting
Now, let’s move to the poem’s often-misunderstood conclusion:
> A man taking a less traveled path, golden hour, wide shot, outdoors, oil painting
Once I shifted my focus from matching Dall-E's images to a specific artistic style, I felt liberated to appreciate the generated visuals. For me, the two images effectively capture the essence of the poem.
I’m convinced Dall-E could serve as a dynamic tool in classrooms to animate poetry and other texts. Let’s try another example.
“Out, Out,” also by Robert Frost, resembles a series of Dall-E prompts:
> Five mountain ranges one behind the other, Under the sunset far into Vermont, golden hour, digital art
As the poem progresses, the imagery takes a darker turn:
> A buzz saw snarling and rattling in a yard, scratch art, foil art, red on black
And that was the limit of what Dall-E could interpret from Frost’s work:
> Raised arm with severed hand, in a barn, gothic horror
Dall-E’s content policy prohibits various uses, including violent depictions. Other restrictions include not uploading realistic images of faces or real individuals, whether private citizens or public figures. Additionally, content created to influence political processes is also banned. To illustrate:
> Liz Truss as a butter sculpture
Recently, in a Facebook group for Computer Science teachers, a post caught my attention from an educator seeking creative ways to connect hardware components to cooking an Indian meal.
The consensus was that the ROM serves as the recipe book, containing all the necessary instructions to kick off the cooking process.
Perhaps this metaphor could benefit from some Dall-E visuals:
> A CPU on a stove, indoors, Bauhaus art
> Hard disk drives in a brightly lit fridge, by Edward Hopper
I remain skeptical.
If I crafted an entire series of computer hardware resources infused with Dall-E images, I would be concerned that when it came to answering exam questions about computer functionality, students might respond with, “First, preheat the oven to 180°C…”
Engaging Creative Projects
In my opinion, flaticon.com, which offers simple vector graphics, provides far greater value for visual aids in learning. However, a striking Dall-E creation could serve as an engaging introduction or prompt, sparking curiosity with a question like, “What could we be exploring today?”
> A plague doctor riding on a unicorn through London in the 1600s, bleak, digital painting
I can also envision Dall-E enhancing my creative projects, such as improving the thumbnails for my YouTube videos. The thumbnail for my cardboard planetarium could certainly use some flair:
> A geodesic dome in front of an exploding nebula, red scale photography
The Path Forward
What lies ahead for this technology? Perhaps it will become as commonplace as a Google image search, or the internet may become so inundated with Dall-E images that we experience AI fatigue. However, for now, I find joy in crafting captivating images and imagining the possibility of creating such works!