With an experimental-making process, I found myself frequently following small ideas in hopes that they would show me a larger truth about what I was creating. I found that in the course of this work, figuring out what should be dropped from the project was also important. This is some of the things that I tried and learned about what this project isn't.
I reached out to a few carnival companies and one person on reddit in pursuit of someone who had practice guessing people's heights in a fair. I was hoping that this information would prove useful in my attempts at making her more of a person. I also thought this might prove interesting content if I ended up going with a podcast, type of a thing.
This was just a step too far away from what I was actually curious about.
This was a difficult one for me to drop. After first failing to connect on my own, I received incredibly helpful guidance from Christian Grewell about a process I could use to get into contact with the flag production companies in China. It is not the most simple thing, especially considering that I do not speak Mandarin. At this stage, I have a fairly comprehensive spreadsheet itemizing about 92 flag brands, tracing them back to the Patent holders, with notations documenting where I'm at in terms of attempts at finding an accurate address. For a few of them, I was able to locate an email and reached out in an attempt to connect over WeChat to no avail.
If I am able to pick this thread back up post-thesis, I fully expect it to yield fruitful results.
Time and effort required to achieve results. I very much would like to continue with this research. This is the biggest hole in my research and I am burning with curiosity about how this image found its way to Chinese flag producers, and how come it came to be so pervasively used.
With a large dataset, looking into utilizing machine learning is a natural inclination. I knew that as an experiment, I wanted to see what it felt like to have a flag photoshopped on to me in real time. I was also very curious about what kind of results would be generated after training a model on this dataset. This is a particularly interesting case because they all utilize the same background, almost as a picture frame.
Later on in the work, there were also opportunities to utilize machine learning in the categorization of the flags. However, due to the earlier experiments, I knew that machine learning ended up feeling wrong for this project. So, I was able to more easily move through conceptual decisions later on because of these early experiments.
For conceptual reasons, I really did not want to subject this woman's image to additional automations. I've tried to be intentional about where, when, and how, I utilize human interference in the story and development of this work. Also, when I tried a few experiments with PoseNet or DCGAN images, I was bored. When I experimented with PoseNet, I found the interaction to be distracting to the point where I didn't even read or look at any of the flags.
As a fairly last-minute idea before presenting thesis work, I was trying to explore the ascii/photo transition moment. As I was building, I was feeling hesitant about just using the photos without intervention. I started to explore different ways of rendering and thought that I might be able to pull the pixel array, omit identifying pixel data, and pass that through a shader. In this way, I wouldn't be hosting the images themselves, just using the image data. Speaking through this with Allison Parrish, helped me realize that this was not the right strategy to solve this problem. However, I did get to do some fun shader play in the process.
I appreciate you choosing a fresh juicy one each every time you show us something new
It was in sharing this that Vince shared his appreciation for my usage of a variety of flags in my explorations, and introduced a rent-free expression into my flag work.
It was a difficult process to use and ultimately did not provide any type of "solve" to what I was looking at. This exploration did help me better define what was bothering me, though. I wanted to think about if someone had access to my entire folder of images, what interventions or intrusions should be there.
With an over-arching task of ethical behavior into this work, I was trying to imagine what would actually be the most helpful to this woman. Prior to speaking with her, I was left imagine the differing ways that she might react and what that could mean for me project. I felt very bad intruding upon this woman's life and letting her know about a problem, without offering up a solution. I realized that one thing I could try to offer up would be to volunteer to swap myself out in place of this photo of her. The work would have to include constructing an equally compelling photograph and then really pursuing the source and location of its pervasiveness. I don't love the idea of my own image being available for photoshops of whatever content, but I could live with it if it meant solving something for someone else.
After speaking with the woman, it did not feel necessary. If I had pursued this idea after the conversation we had, I think it would have been more of a gesture related to my own ego, then it would have been about this person's actual needs and feeling.
Along a similar vein to the above, I was trying to think about interventions into the usage of this person's image. Assuming that the bulk of these edits are happening in photoshop (an utterly unfounded assumption!), I thought this might be an appropriate form. Photoshop already has content interventions in the case of counterfeiting money. I think there is something interesting about putting a tremendous amount of effort into preventing the usage of this one particular photo. I think this would have been a statement about the difficulty/futility of trying to rid something from the internet, especially when it has been reproduced at such a massive scale.
Theoretically, this could have been a singular solution that scales up. Building this type of intervention tool one-image-at-a-time, could lead to a submission/database type of format. People could submit images of themselves (this is undercooked and needs an aggressive amount more ethical think throughs!) that they don't want people to photoshop. Then photoshop users would have to intentionally opt in to a tool which restricts them. Would most likely be a performative gesture, rather than a practical application.
This would have likely been continued if the woman had felt differently about her image being used or about my involvement in her storyline.