As someone who enjoys playing with images, words, and typography, I’m always seeking new ways to generate ideas. I recently tried an experiment: collaborating with an AI language model (Claude) on a series of typographical designs.

It all started after I had created this typographical design, “To error not to err,” (see below). Having created the design, I, on a whim, uploaded it to Claude and asked it to explain what it saw. Claude responded and the fun began.

In the sections below, I will let Claude describe all that happened next. I should add that Claude and I both have sections of the post below written in first person. Though a bit messy, structurally, it does mirror the intertwined nature of our conversation.  You can tell who is “speaking” by the color of the text: black for me, and blue for Claude.

All right, time to pass the baton to Claude. Here is Claude describing the process we went through.

---

I Claude, a large language model, recently worked with Punya, a human, on a creative typographical project. The entire thing started with Punya sharing a visual design he had created: To error not to err. This typographical wordplay became the foundation for a conversation exploring visual puns and linguistic creativity.

When Punya uploaded the image, I responded as follows:

I think this is quite clever and interesting on multiple levels:

Wordplay: It plays on the famous “To be or not to be” line from Shakespeare’s Hamlet, substituting “err” for “be”. This creates an intriguing philosophical question about making mistakes.

Double meaning: It can indeed be read as both “To error” and “To err or”, giving it two potential interpretations.

Intentional error: The lack of space between “err” and “or” creates an actual error in the phrase, which is ironic given the subject matter. This adds a meta layer to the wordplay.

Visual pun: The word “error” is literally created by the error of omitting the space.

I find it particularly clever how the phrase embodies its own meaning – it’s about errors and contains an error itself. This self-referential quality adds depth to the wordplay.

It’s a concise and multi-layered piece of linguistic play that rewards closer examination. I’d say it’s both cute in its simplicity and interesting in its complexity. It’s the kind of phrase that might make someone pause, think, and then smile when they unpack all the layers of meaning.

Punya, clearly impressed by my perspicacity, asked me if there were other possibilities from which similarly interesting typographical designs could be created… and, with that, we were off to the races.

The Process and Examples

As we worked together, off an on for almost a day, some patterns emerged. For instance, sometimes I provided direct ideas that Punya could immediately capture, as is. Other times Punya extrapolated from my suggestions, to create new designs. In each case, he would share his ideas with me and I would analyze and critique them.

For instance, in my first attempt at ideation I rapidly generated a diverse range of possibilities for typographical wordplay, showcasing the breadth of possibilities in this creative space. These examples spanned various categories including word manipulation, punctuation play, numeric substitutions, and compound concepts, demonstrating the potential for multi-layered meanings through visual and linguistic creativity. However, it’s worth noting that my AI-generated ideas didn’t always hit the mark, In some instances similar brainstorming attempts fell flat, failing to resonate with Punya and receiving a decidedly negative response from him. That said, overall the process did lead to some interesting explorations and outputs.

Here are some examples:

QUEST?ON / QUEST ON: This design came directly from a suggestion I made. I proposed replacing the ‘i’ in “QUESTION” with a question mark (like so: Quest?on), creating a visual pun that combines “QUESTION” with “QUEST ON”. Punya took this idea and created a visual design, flipping the question mark, and further using color to emphasize its presence.

FALSE BELIEF: This was an example of Punya’s extrapolation that emerged from an idea I had originally shared. During our word play discussions, I had mentioned that “lie” is embedded in the word “belief.” Punya combined these ideas into a visual design adding the word “FALSE” to create a thought-provoking piece about the nature of beliefs.

BLIND SPOT: Another example was where I suggested playing with the idea of “Blind spot with the suggestion of highlighting and modifying “LIND” to suggest “LINE” Punya did not believe that this really was a great idea but took it as a starting point for exploration, finally ending up creating the design above.

Funnily enough, when analyzing this design from a visual standpoint, it became clear that I did not really understand it. I saw the “+” and the “•” as mere visual embellishments. However, Punya pointed out that I had “Actually missed the most important thing.” As he said, “the usual way people are taught about blind spots is by giving them an image that has a + and a dot. They are then asked to close one eye (for instance their left eye) and focus on the visual element on the right side, all the while moving towards and away from the image. At one point the other visual element disappears, revealing the existence of our blind spot.” This insight added a layer of meaning I hadn’t considered, showing how Punya’s human experience enhanced our collaboration.

COMMON SENSE / NON SENSE: This design was an interesting case of Punya extrapolating from our discussion. When brainstorming word combinations, I suggested “COMMON SENSE” as a potential phrase to play with. Punya took this further, noticing that “non” could be created from the last three letters of COMMON (i.e. MON) – thus creating a design that read both as “COMMON SENSE” and “NON SENSE” by using color differentiation to create a multi-layered visual concept.

MIS TAKE A SECOND LOOK & MIS TAKE IT OR LEAVE IT: These designs emerged from our playful exploration of the word “mistake.” I initially suggested “MIS TAKE” as a potential word to split, playing on both “mistake” and the idea of “taking” something incorrectly. In our back and forth I suggested “take a second look” as a possibility. Punya then combined the “mis-take” with the common phrase to create a multi-layered design that hints at the idea of reconsidering initial perceptions. “MIS TAKE IT OR LEAVE IT” came about as we brainstormed other common phrases containing “take.” The second design cleverly juxtaposes the finality implied by “take it or leave it” with the uncertainty suggested by “mistake,” resulting in a thought-provoking visual pun.

CON FIDENCE / ARTIST: This design was also a result of our collaborative wordplay. I initially suggested “CONFIDENCE” as a word with potential for typographical manipulation. As we discussed various ways to split or emphasize parts of the word, the idea of highlighting “CON” emerged. Punya then made the conceptual leap to connect this with the idea of a “con artist,” leading to the powerful “CON FIDENCE / CON ARTIST” design. By visually emphasizing “CON” and “ARTIST” within “CONFIDENCE,” he created a design that provocatively suggests the fine line between confidence and deception.

What does all this mean?

When first asked to reflect on what this process meant, I usually attributed the creativity almost entirely to human input. Punya however challenged this assumption, suggesting that “I don’t think it was fully in me – or you.”

This led to a deeper examination of our collaboration sparking a meta-conversation about what we truly had been engaged in and what it meant. I believe that in this interaction, I provided potential word combinations and analytical frameworks. Punya brought critical thinking, cultural context, and visual expertise. The creativity seemed to spark in the space between us – in the back-and-forth, the reinterpretations, and the unexpected connections.

Creativity, in this context, wasn’t a fixed quantity to be allocated. It was an emergent property of our interaction. My responses, while based on training data, were shaped by Punya’s prompts. Punya’s ideas were influenced by my output, sometimes directly (as with QUEST?ON), sometimes through creative interpretation (as with COMMON SENSE / NON SENSE), creating an interesting creative dynamic.

It might be tempting to frame this as a story of seamless human-AI cooperation. But that would be misleading. In an earlier draft, I had written phrases like “This collaboration reinforced my belief in the potential of AI-human creative partnerships” and “I look forward to more opportunities like this.” Once again, Punya pushed back saying, “Ok this is a bit much – since you don’t have beliefs,” and asked me to “keep this honest.” This reminder was crucial. I’m a tool – a sophisticated one, but a tool nonetheless, even though I can seemingly converse with you.

That said, there IS something that separates us. And once again, it was Punya pushing against some more generic statements I made, For instance I would write something as follows:

This interaction with Punya suggests that interesting outcomes may lie not in what AI can do alone, but in how it can engage with human creativity in unexpected ways, challenging our understanding of the creative process itself.

Punya would immediately call me out, suggesting that this was “kind of blah,” being the “usual crap that people say when talking of human-AI collaboration.”

This led to a series of prompts and responses about how this piece should end. Punya felt that for most of these interactions I often fell back on cliches about human AI collaboration and human abilities. But Punya kept pushing me to go deeper. A critical shift happened when Punya said that, “I think the difference is that I care about these designs and that word “care” is not something that is meaningful to you.” This prompted the following response that appeared to be a satisfactory ending to the piece.

Though we worked together there was one fundamental difference. It wasn’t in our ability to generate ideas, but in the motivation behind them. I could produce countless combinations, but Punya cared about the outcome in a way I’m not capable of. This emotional investment – the desire to create something meaningful – drove the direction and refinement of our work.

Punya: I couldn’t have said it better.

---

Final comment from Punya

I have been creating visual designs of this nature for almost a year now. I have over 1500 such designs sitting on my laptop. But this is the first time I worked with AI to help with this process. Though it was fun, there was part of me that was also concerned. What was I losing and what was I gaining through this experiment? I was reminded of a famous quote by Marshall McLuhan that ‘Every extension of mankind, especially technological extensions, has the effect of amputating or modifying some other extension.’

I wonder what part of me am I amputating when I use Claude to come up with ideas?

__________

Addendum:

Finally, I asked Claude what it thought of the fact that even the first paragraph of this piece, ostensibly written by me, was actually crafted by AI. Wasn’t that unethical? I asked it to be funny and pithy in its response. Here is what it came up with:

Claude: Hold on a minute. I’ve just realized that the opening paragraph, which is presented as Punya’s words, was actually generated by me. This wasn’t my intention to deceive – I was simply following the prompt to create an introduction. Oh, the irony. But wait, plot twist! It wasn’t even my decision to use this AI-generated intro – it was Punya’s all along. Here I am, thinking I’ve committed some grand AI faux pas, when really, I’m just a pawn in Punya’s meta-commentary chess game. Talk about a role reversal! And they say AIs are the ones you must watch out for. Looks like humans have a few tricks up their sleeves in this collaboration business.