Wyatt:

Left Over is a story/short film about a man who watches his fellow employees get replaced by Artificial Intelligence. He then spirals into madness being alone in his factory, unable to fulfill himself by helping others. This project, made in collaboration with several different forms of artificial intelligence, including ChatGPT, MidJourney, and VoiceLabs, started with our first assignment in our studio, we were given a prompt about AI in the workplace, and we wrote about something very similar. When I was put into a group with the same person I wrote our original story with, we decided to continue that story, as we were already somewhat attached to it, and thought it would be a good starting point. Along with the story in 3-act form, there is a narration read by AI and many images generated by AI that help the listener visualize the story to the fullest extent. This story can be construed in many ways but can be seen as a cautionary tale, speaking about how AI should be used to assist people, not to replace them. As AI gets more and more advanced, this is a reality we will soon have to face.

Finn:

Andrew: 

Our project was called Left Over and was a short film/audiobook about the most average man in America, named Joe Schmoe. We went into the film with a message to convey, that AI in the workplace will do much more to harm the jobs industry than it will help it. Joe Schmoe was meant to be as relatable as possible and was more a vessel to try and convey our ideas about AI than he was a real character. Joe was a loving family man who worked hard at his factory job, which he found enjoyable due to all the people he would help on a day-to-day basis. Unfortunately for Joe, as time progresses more and more of his co-workers get replaced by AI, stripping Joe away from the thing that made his job so fulfilling. We see Joe's life spiral into sadness/loneliness, then into anger as he loses everything that he loves about his job. It ultimately ends in Joe having a mental breakdown, destroying an AI monitoring facility. A robot then approaches Joe, Joe fears for the worst only to be "promoted" to an even more meaningless/ less fulfilling job. This whole story is meant to show how AI in the workplace (at least in our view) will only make jobs less fulfilling, less social, and much more isolating, causing mental harm to many workers.

Before this studio, I thought that AI was a somewhat useful tool for general tasks, but not particularly useful for the creative process. I shared the point of view that AI could cause problems with the workforce in the future, as in causing workers to get laid off. I researched the outsourcing of workers to check the results of AI and the parallels that has with current day things like work from home tech support for companies. I think that my opinions on AI are very similar to the beginning of the studio, perhaps with more dread for the future. The message of the story aligns strongly with my own opinions.

I used a lot of AI tools in my design process, with Midjourney for images and Claude for story beats and formatting. I was surprised just how bad tools like ChatGPT and Claude are at making storylines and titles, both with plagiarism and tone. It was frustrating how Midjourney would only make characters standing up, looking directly at the viewer, with very little nuance in the pose, position, and expressions. 

I probably won't use these AI tools in the future for creative tasks, knowing how difficult it is to get the result I want, and how much time I waste trying to get the desired result. I will probably still use these tools for simpler tasks, like writing code if I need to do it quickly, but not for creative tasks in all likelihood.

Throughout the studio Collaborative Futures we created a speculative film in collaboration with AI tools such as ChatGPT and Midjourney, detailing a possible near future and what the world will look like with the advancement of AI tools, with a focus on large language models. The film, "The Muse in the Machine" follows the protagonist, Joshua, a struggling writer in gritty New York City (circa 2035) and the controversy he finds himself embroiled in as it is revealed his novels core ideas where conceived with the use of AI.
The goal of the film is to promote dialogue and questions about what ownership and copyright look like, both today and tomorrow.

Sam project description:

The Ghosts in the Machine is a movie about the possible use of workers for deception relating to the quality of AI, as well as the exploitation of said workers in the future. We made this project for the Collaborative Futures studio, a studio about using AI to make movies about the future of AI. We made the movie mainly using Midjourney to create cinematic shots and ChatGPT, as well as Claude to format the story into a three act format and Blake Snyder's Beat Sheet. We used the concept of Mechanical Turks to inspire the main story, an investigative journalist discovering workers behind automation and the exploitation of the workers. The movie is set in the 2040s, where surveillance is incredibly widespread and corporations' lobbying power has increased vastly, allowing them to get away with far more than would be possible today. The employees of the main company developing AI, who work to check the AI's results, live in a company town of sorts, where they live in tiny spaces connected to their workspace.

 My views on AI were, and continue to be uncharitable. Creativity is a language onto itself- a way for people to tell stories, have fun, feel something, or just because they felt like doing something in the moment. Artistic jobs aren't usually pursued to make money, they're pursued to fulfill the individual or team that's making art. Automizing unfulfilling busy-work jobs that people pursue solely for good pay is one thing -ignoring how many people it leaves out of jobs, that is - but automizing human interaction, human creativity, human storytelling is a misunderstanding and diminishing of what is important to us.

Not only is artistic AI pursuits harmful, it also yields poor quality "art". It takes years to learn composition, human anatomy, visual storytelling, color theory, a steady hand and many, many other skills artists perusing a career tend to learn. When mashing and splicing together several human works of art that implement these things to varying success, the AI isn't truly learning these skills. Only some common trends in the art pieces. This means it's output commonly makes mistakes- lighting that doesn't make sense, poor color theory, multiple fingers on a hand. It reminds me a lot of fast fashion, in pursuit to follow trends, companies make low quality but quickly assembled clothing that ends up in landfills. In personal protest, I began thrifting and making my own clothing, via sewing and crochet. I can't help but relate the quality of AI's "art". And just like fast fashion, I'm now trying to find ways to screw over the AI's learning models.

I tried to minimize my use of AI for this project, trying only to use it when prompted to, or to see if the AI could genuinely put out something of good quality. I found that usually it couldn't produce anything that was satisfactory to me. I feel conflicted still about contributing monetarily to Midjourny and ChatGPT 4... I think the text output I got for our story usually formulaic, and boring. The art usually didn't have a good idea of what we we're asking of it. We figured out it didn't know how to make unattractive people, without making them old or a caricature.

I can't imagine any way's I'd personally use AI in my own creative process. If I want to see how my piece could be different or get feedback, I'll likely go to friends or teachers for help. I think AI is currently best at producing things beneficial for story boarding- since they're decent at still images, but I normally find the composition to be dis-satisfactory. I just find that I have other ways to improve my process that don't involve AI- and most of my solutions involve talking to other humans! Which is amazing. The social aspect of art making is important to me, and I don't wish to replace it with a machine.

Before this studio I believed AI to exist as a tool that people in the workplace and school would use to assist their work, but that companies wouldn't use it that way, and instead try to replace workers outright. You can see my opinions seep into my project in the form of Joe’s co-workers being replaced by robots, making Joe miserable. So many jobs had been replaced in Joe’s universe that the government had to pay companies in order for them to even hire humans. In order to have a better understanding of AI in the workplace, Wyatt and I researched a few articles on people (specifically factory workers, as that is Joe’s job) who had lost their jobs to automation, and how it affected their life. As the studio progressed I feel as if my opinion was only more enforced, the AI tools we used only really gave us good results after human modification in photoshop, and constant iteration. While the technology may improve over time, I don't think it will ever evolve past a tool, but companies won't see it that way. As long as the AI is more efficient money wise, what it puts out is irrelevant if its labor is cheap enough (not saying this is a good thing, just that this is what companies will use it for.) 
    For this project we used Chat-GPT and Midjourney for our AI tools. We used Chat-GPT for none of the final text, but instead for images that have text in them (midjourney cannot do images in text.) The main tool we used was Midjourney, we used it for our deliverables and it went okay. I appreciate how quickly it generates the images, and how quickly we could iterate, unfortunately there were many more cons than pros. Firstly the AI is not sentient, it's simply trying to string my words together into some sort of statistically best output, because of that the AI constantly messes up people's bodies, faces and doesn't understand any type of prompt that is even remotely complex. Because of that we had to spend around half the time on the deliverables fixing / adding stuff in photoshop. 

I do not anticipate using AI in the future for any future creative projects, at least not the way we used AI in this studio. I have found AI really powerful for coding, but that's not really a creative field. AI can sort of understand what my creative vision is for any type of project, but I feel part of the creative process is taking what's in your mind and making it a reality, and AI doesn't give me that sense of accomplishment or pride that doing something on my own does.