Today In class, Clio and I experimented with creating different prompts for two different large language models: ChatGPT and Google AI’s Gemini. We wanted to use these two as they are two of the most popular AI LLM companies. We did two separate ones for the sake of comparison.

PartI:

I experimented with giving It the persona of having the voice/style of Sylvia Plath and writing a poem about a relationship. We found that Clio’s Gemini output was not only generic and not very ‘good’ but it didn’t really sound like Plath. She for one, does not rhyme in her works, however Gemini outputted a completely rhyme-bound piece.

My ChatGPT out put however was more promising, with words, phrasing, and style much more similar to Plath’s works. It still had flaws in genericness and some inaccuracies in how she would write, and some of the themes/flow did not make much sense and sounded like it was “fake deep/sad,” but it was significantly better. We think this is because I used the ChatGPT customization tool in which I gave it specific personality traits like being “poetic.”

Part II:

We then experimented with asking it to write short stories and song verses.

For the short stories we asked it to write “a short story about the feeling of summer.” For the stories we again felt that my ChatGPT output was a much ‘better’ story while the Gemini one was a little odd and bland. It was interesting to see that mine was extremely romantic, soft, deep, and nice sounding.

Arriagada use terms like novel and valuable to define creativity. This is true and interesting, especially considering everything I learned this week—specifically around how AI-generated art gives off a sense of emptiness. I would now define creativity as the unique ability to move the audience with one’s work.If I had to challenge any part of the definition, it would be the new aspect. In class, we spoke about how even Shakespeare’s work was drawn from other traditions, and how rock music was drawn from blues. My mother always says, “There’s nothing new under the sun,” which means everything draws on inspiration from the old or is a manipulation of it especially nowadays.

I asked Perplexity to “write a poem that rhymes,” a very generic prompt, and it produced A Journey Beneath the Sky. Then I asked it to take that same poem and, using Shakespeare’s voice, rewrite it as a sonnet.

I really do not know what to say about this because I feel like elements of the original poem were lost, and I ended up with a completely different poem in sonnet form, full of “thy” and “thou.” Interestingly, the second time around, it explained how the sonnet structure “mirrors the Bard’s themes of cosmic wonder, mortal limitation, and the interplay of light and shadow.”

To me, this is unsatisfactory. I think AI can be creative, but it requires extensive prompting—and even then, it feels soulless. It can be used as a tool, not the author and finisher of an entire artwork

Boden, Margaret A. “Creativity and Artificial Intelligence.” Artificial Intelligence, vol. 103, no. 1, Aug. 1998, pp. 347–56, https://doi.org/10.1016/S0004-3702(98)00055-1.

Arriagada, Leonardo, and Gabriela Arriagada-Bruneau. “Ai?S Role in Creative Processes: A Functionalist Approach.” Odradek. Studies in Philosophy of Literature, Aesthetics, and New Media Theories, vol. 8, no. 1, 2022, pp. 77–110, https://philarchive.org/rec/ARRARI-2.

This week’s dive into AI creativity and the readings on it were interesting. One particular prompt regarding its capacity to create poems caught my attention. I was stunned by how the LLM balanced traditional poetic structure with unexpected twists. This challenged by assumptions about creativity.

To me, creativity has always been a blend between discipline and spontaneity, order and chaos. To reimagine a machine form something as linguistically complex that captures this version of creativity was truly mind-boggling Here’s the AI’s response regarding this:

“In the quiet library of midnight,
the clock whispered forgotten secrets
while shadows danced on silver pages,
each step a quiet rebellion against the dawn.”

The interplay between structure and surprise stood out to me. There is an element of traditional poetic framework coupled with an unexpected energy. This mixture of predictability with delightful unpredictability is what I consider the essence of all creativity. This resonates with the philosophical perspectives we discussed, which suggest that creativity involves recontextualizing the familiar to reveal deeper insights (Loi et al., 2020).

The paper we read also raised important ethical questions. As AI systems become more integrated into creative fields, we need to consider the moral costs of relying on these systems. For instance, when mainstream AI stifles the kind of open-ended exploration that fuels genuine creativity, there’s a risk that we might lose something irreplaceable about human ingenuity. In my view, while the AI-generated poem may not represent a radical breakthrough in originality, it does prompt us to think about how we value and attribute creativity, particularly when the creator is not human.

In this week’s research, I found learning about the inputs and outputs of prompting LLMs to be very intriguing. Specifically, the Persona Pattern (Prompt Improvement) made me interested to learn more about the process of large language model prompt engineering and how giving artificial intelligence more characterization can help reach more in-depth and accurate responses.

To utilize the Persona Pattern prompt, I wanted to experiment specifically with the LLM: ChatGPT. I decided to used ChatGPT because it is customizable to impersonate specific personas. For this experiment I decided to customize it with a few general personality traits, one of which being “talking like a member of Gen-Z.”

To help it, and for the sake of the experiment I wanted to take on a little bit of the persona myself to engage more with it. I thought that experimenting with a typical young-adult romantic topic would be interesting in funny. I asked it if I should ‘like totally dump my bf?.” I received an almost comical and slightly ridiculously gen-z slang-saturated response, with some good advice and a little bit of humor. It seemed really engaged and I found it interesting that every time I generated a prompt, it would ask something like “need any more like totally fire advice?” and start generating even more things I could ask it to do.

Shen Article:

During our research lab and this week’s findings, media articles, etc I found a lot of interesting takes and facts on ethics in the world of artificial intelligence and large language models. 

                  I thought it was incredibly interesting to learn about how different nations and cultures use large language models. 

This lab specifically prompted my idea for my final project as it made me question the ethics behind large language models and artificial intelligence in regards to the exploitation of other cultures and how it directly benefits first world consumers such as ourseleves. 

It begs the question: at what cost is our learning/creativity/functions through large language models and artificial intelligence, reaping those who carry these technologies on their backs (considerably underpaid). Additionally, it is important to note the accessibility of these clearly differ from around the globe. 

In our group discussion, we referenced specifically Chinese culture and how their access to these technologies is seen as beneficial to many of the nation due to China’s industrial culture. Their nation prides itself on becoming a technological superpower and exceeding standards of technological safety/security.

Here in the united states, from a privledged, first-world stand point, many citizizens here see LLMs and AI as dangerous and threatending to creative intergrity, jobs and industry, education, and more. This differs from the majority of the Chineese perspective as they see benfits in surveillance control, facial recognition, genetic profiling, and other forms of government-oriented safety measures for their country. 

It is also important to note the disadvantages that were also discussed in the article such as social control. The “Big Brother is Always Watching” effect being unforgiving to certain artistic and political freedoms of the citizens. 

I also found an article from the Journal of International and Public Affairs from 2019 that discusses the certain benefits in 3rd world countries that AI has implemented such as structured learning methods, easy money transferring models, etc.

I used the flip interaction for the first time. I have heard of it before in a previous digital studies class. So I knew the concept of it, but I also didn’t know how it would work out. My prompting could be worked on however I tried to give as much details as possible for the LLM. I used google’s Gemini, which can generate answers and pictures. I needed the picture function, because I wanted to design a kitchen with AI and I wanted to see the end result.

My starting prompt was:
you’re a kitchen designer, I want to remodel my kitchen which is 4 meters long, 3 meters wide and 2.5 meter tall. on the south side there is a big arch which covers the whole wall. Right next to it on the east side there is a doorway to the outside and exactly opposite there is a door to a staircase. on the eastside there are two windows. I want you to ask me questions about the design of the kitchen and if you have enough information I want you to describe and then generate an image based on my answers
I then quickly corrected the AI, because it asked my a bunch of questions at the same time and I wanted it to ask one question at a time.
The prompt could definitely use some work, A prompt can influencesubsequent interactions with—and output generated from—an LLM by providing specific rules and guidelines for an LLMconversation with a set of initial rules. (White et al, 2023). So by tweaking the initial prompt to be more specific, in what I expect from the LLM I could get better result.

This type of interaction with LLM is useful in some case. As per other stuff with AI I think it is a great starting point. If you have no clue on how you want to redesign the kitchen, AI can give you some things to think about. However when you go further into the process the AI become unreliable and starts to create things which are not correct. As you can see in the starting prompt the arch is not on the same wall as the window, however in the AI generated picture it is generated on the same wall.