Thesis Documentation
AI Log
September 11 – December 4 •. AI Log
Week 2 – Initial Investigation
ChatGPT Prompt 01: With this first prompt, I wanted to learn how visual communication design can challenge the idea of glass children sharing their thoughts and feelings to others. The answers they gave me were good, however, I know I can expand on them when I start to expand on my research in the coming weeks and start the design process.
ChatGPT Prompt 02: For the second prompt, I wanted it to suggest to me different scenarios in a design format in today’s world that may assist with my ideation. Unfortunately, it gave me different forms, which was helpful, however, I wasn’t necessarily ready to look through the format just yet. So, I ended up creating another prompt as a follow up to explain what I had envisioned.
ChatGPT Prompt 03: For the third prompt, I used it to follow up with my second prompt as it didn’t answer what I was looking for. I ended up switching the term from ‘design format’ to ‘design language’ which I think definitely helped me more than the last prompt. Some of these ideas were interesting, especially “silence vs. voice”, as I can tie it into overthinking somehow. It could be interesting to explore as I move further along the project.
ChatGPT Prompt 04: Finally, with my last prompt, I wanted it to give me examples on how I can bridge the gap between communicating with glass children and to outsiders. Some of the answers ChatGPT gave me were interesting. I would like to integrate some of the ideas it has given me and expand on them later on.
Week 3 – Ideas & Sketches
Within my process, I used Adobe Firefly to help me explore different ways to explore how to show guilt and invisibility within my thesis project, and the results they gave me were ultimately very interesting.
Prompt 01: “I'm doing my visual communication design thesis on glass child syndrome. My question is, "How can I use visual communication design to validate, and reveal the invisible thoughts & feelings of a glass child?" and my argument is, "By incorporating sources of nostalgia, such as photo albums, home videos, and my personal experience as a glass child, I aim to create a space where glass children can see a story reflected in their own experience and where others can learn and connect with my story." I'm looking for some visualizations as a starting point for sketches and to help me build concepts for my thesis project. Would you be able to create a visual with the idea of nostalgia, and the themes of guilt and invisibility which are the core themes I'm looking into.”
Response: Ultimately, these iterations were unsuccessful—they were too literal, and didn’t evoke the emotional aspect that I was looking for. I knew that I had to train the system to make my visuals seem less obvious.
Prompt 02: “I'm not a fan of any of these. I'm more so looking for something dynamic. Think sources of nostalgia such as photo albums, and home videos. Maybe be a bit more dramatic?”
Response: While these visuals were better, they didn’t invoke what I was truly looking for. I felt that the system focused more on nostalgia rather than the feelings on guilt and invisibility, which ultimately was my mistake as I didn’t include it in the brief.
Prompt 03: “No, no. Use this style as an example, but think of nostalgia, photo albums, and home videos. Imagine that concept with the misunderstandings of glass child syndrome (I should have specified that it is the sibling of someone who is disabled).”
Response: I gathered an image from my moodboard as a reference in order for Firefly to better understand my concept, where I wanted my imagery to be more faded (to represent invisibility). While the visuals did look better, I still realized that I wanted my imagery to be a bit more abstract, and not directly obvious that my main ideas are guilt and invisibility.
Prompt 04: “Don't really include portraits. I'm looking for visual elements that symbolize guilt, and invisibility. also include faux imagery of childhood memories and, include typography in the visuals as I plan to use type as a way to symbolize my thoughts.”
Response: These results came out incredibly interesting, especially the middle two. I found those to be the strongest, and although the software didn’t include typography, I’m glad because the software gave me two strong images that relate to feeling invisible, and guilt that I can reference when creating my pieces.
Week 5 – Design Drafts: First Review
This week, I used AI to assist me with rephrasing my thesis proposition. My current thesis proposition was a little bit wordy, so my goal was to use ChatGPT as a resource for me to assist me in coming up with a revised proposition that aligns with my thesis question, and initial argument, as of now.
Prompt 01 Response: With this prompt, I wanted to emphasize to ChatGPT that I wanted to make it more succinct and to make my proposition make more sense. The results it gave me were great starting points that I took advantage of when rewriting my proposition.
Prompt 02 Response: Although the results the first prompt gave me were good, I knew I wanted to dig a little deeper, so I attached my thesis question I created earlier in the semester. Using that along with my current argument benefited ChatGPT by combining the two into one proposition.
Prompt 03 Response: Once I finished writing a draft for my proposition, I sent it to ChatGPT to give me feedback—I ensured to emphasize feedback as I didn’t want to give me solutions. It gave me a suggestion that I should make my proposition more dense.
Prompt 04 Response: I then asked ChatGPT to elaborate a bit more on how I can make it more dense. I got the clarification I needed when I realized that there were too many clauses in one sentence that it was best to break it down.
Prompt 05 Response: Once I improved my proposition, I asked ChatGPT to read it through to ensure it sounded better. It gave me the clarification I needed that adding another sentence to my proposition was a smart decision.
Reflection: I felt that this week, ChatGPT really helped me to ensure that my proposition made sense and to fix any grammatical errors. It was a great resource for me because the clarification it gave me helped me get more confident with my writing.
Week 5 – Revision 1
This week, I explored Generative Fill on Photoshop and see how it can work with image manipulation and see if it is a viable choice in the future. I took one of the photos I planned on using in my book and tried simple prompts where I can content aware areas in the photo..
Prompt 01: Can you expand the sky to the bleed of the page using content aware?
Reflection: I felt that generative fill did a good job with responding to the prompt. While the second result included type on top which was not what I asked, result one and three did a good job of reflecting that—I chose result one to continue with this experiment.
Prompt 02: Can you expand the sand to the bleed of the page?
Reflection: While the first prompt did a good job with following what I asked, the second prompt, not so much. It included type which was not what I had asked for, and the execution was not done properly.
Prompt 03: Can you expand the sand to the bleed of the page using content aware?
Reflection: The mistake I used last time was not including the words content aware in the prompt. However, it still didn’t give me what I was asking for—I ended up using content aware to fill the sand.
Week 6 – Revision 2
This week, I didn’t use AI much to assist me with visual development. However, I used ChatGPT to assist me with coming up with ideas on how to make the pages of my book feel cohesive.
ChatGPT Prompt 01: With this prompt, I asked ChatGPT to assist me with thinking about ways I can ensure cohesion with effective sequencing. The ideas they gave me helped a lot, but I was looking for more specific answers on how I can use my design and layout to enhance the sequence.
ChatGPT Prompt 02: With this prompt, I asked ChatGPT to be more specific on how I can use paper to think about sequencing. These were incredibly helpful as it gave some ideas on how I can think beyond the traditional book, and while I was already integrating ideas such as vellum and fold outs, I would like to look into some of the ideas they were giving me even further.
ChatGPT Prompt 03: With this prompt, I asked ChatGPT to generate an image on different folding techniques for my book. While obviously the images used in the generation aren’t images I want to use, some of these were helpful, but not fully. I knew I had to ask ChatGPT to be more specific.
ChatGPT Prompt 04: With this prompt, I asked ChatGPT on how I can expand the use of vellum or transparent paper within my project. The ideas they gave me were helpful, however I wanted to see what it could generate so I could expand on it when I started designing my product.
ChatGPT Prompt 05: I asked ChatGPT to generate an image on how I can use vellum. The results were not what I was looking for. They were too clean, and didn’t reflect what I was looking for, which was using cutout elements of an image and printing it onto the vellum.
ChatGPT Prompt 06: I ended up revising my wording for ChatGPT to see what else it could come up with, but I still wasn’t impressed. The images they used were too blocky, and I felt that it would be more beneficial for me to come up with the imagery myself.
Week 7-9 – Design Production 1-3
Throughout the production stage, my usage of AI was extremely limited. I chose not to use AI because I felt that relying on another point of view (even if it was ChatGPT), would ultimately lose the idea that this thesis was my story from my point of view, and using AI would feel like an injustice. I was worried that I would lose the human aspect by using AI throughout the production stage, and I felt that AI helped me in developing concepts, rather than design production. The design production stage was a chance for me to showcase what I have learned throughout my time in the program.
Week 14 – Further Findings
For this phase, I used ChatGPT as a way to assist me on how I move forward based on the material I have uncovered. I asked AI prompts based on the visual research I had uncovered.
ChatGPT Prompt 01: I found that ChatGPT’s response to my prompt was very informative. It gave me a good insight on how I can use Wilkins’ and Fella’s work as inspiration for my project to enhance my imagery in the way that would be beneficial to my project.
ChatGPT Prompt 02: I found that my follow up question to ChatGPT was very useful. It gave me design movements to think about and ultimately would be of benefit to my project if I chose to research them further.