Module 3 of the course delves into the fascinating world of AI narratives. It offers a comprehensive overview of AI in writing, showcasing examples of AI-generated fiction and its impact on readers. This module encourages students to explore the unique capabilities of AI in creating narratives that differ from traditional print stories, examining the potential of AI to transform storytelling.
Learning Outcomes:
- Gain a broad understanding of AI's role in writing and storytelling.
- Evaluate AI-generated writing alongside traditional human-crafted stories.
- Analyze and contrast the characteristics of print narratives with those of AI-generated content, focusing on storytelling differences, audience engagement, and publication possibilities.
Wu, J., et al. (2021). "Recursively summarising books with human feedback."
This groundbreaking study explores the use of AI in summarizing complex literary works. The paper delves into the process where AI algorithms are trained with human feedback to create concise, yet comprehensive summaries of books. It presents a unique intersection of human creativity and AI efficiency, offering insights into how AI could augment human understanding of literature.Novel AI (2021).
This website serves as a platform for AI-powered writing tools. It demonstrates the practical applications of AI in creative writing, providing users with AI-assisted tools for crafting narratives. The site exemplifies the blend of AI technology with human creativity, offering a glimpse into the future of storytelling where AI plays a collaborative role.Benson, D. (2022). "AI in Fiction and the Future of War."
David Benson's article probes into the use of AI in writing war fiction, examining how AI can generate narratives that speculate on future warfare. The piece highlights the capacity of AI to extend beyond current human imagination, offering new dimensions in fiction writing, particularly in speculative and science fiction genres.Chubb, J., et al. (2022). "Expert views about missing AI narratives: is there an AI story crisis?"
This paper delves into the discourse around AI narratives, addressing concerns about potential gaps in AI storytelling. It raises critical questions about the representation and diversity in AI-generated content, examining whether AI can truly capture the breadth and depth of human experiences and storytelling diversity.Howlarium (2022). "What will AI do to reading?"
This article explores the implications of AI on the future of reading. It discusses how AI might transform readers' experiences, from interactive storytelling to personalized narrative journeys. The piece contemplates the changing landscape of reading in the age of AI, probing the potential shifts in reader engagement and content consumption.
See our eClass for details on assignments or course work.
The Wu, J., et al. (2021) article provides insight into how AI training models are used to create narrative summaries. The models described use behavioral cloning (BC) and reinforcement learning (RL) to create summaries that are comparable to human written summaries. While this concept could eventually save time creating summaries, I do wonder what the intrinsic value is and what we might lose if we pursue this over learning to summarize text ourselves. Benson, D. (2022) examines how AI is addressed in fiction, and how fiction can be used to “explore the possibilities and risks of AI” (para. 4). These risks and possibilities are often impacted by the culture of the reader, or the reader’s first encounter with AI in narratives. However, science fiction is often a precursor to science fact, used to inspire human action and development. How will these individual perceptions of AI from fiction impact what is created in actuality?
ReplyDeleteKeeping in mind that human authorship inspires creativity in entrepreneurship, what the advent of AI authors mean for the future scope of human development? Novel AI (2021) is a powerful writing tool that creates a story with very little prompt from the user. To do this, it requires a large dataset to learn from. As Chubb, J., et al (2022), these datasets are often biased, with gaps in data from different genders and backgrounds. Even if the dataset grows based on user inputs, and becomes more diverse, what does this mean for human creativity? If we begin to utilize AI for use in the creative industries to save time or ease the process, will we continue to grow creatively? If we don’t exercise our brain, will we lose the ability to explore new possibilities? AI has huge potential, but I think we need to look carefully at both the risks and possibilities. AI should be used to free up time to allow the human mind to explore and grow, not replace the fundamental actions that would support that development.
This module has me incredibly excited about AI tools, and I appreciate the opportunity to discover and learn more about tools such as RYTR, Novel.ai, and Microsoft Image Creator. I am reminded of the plot line in the movie “Hidden Figures”, where a worker (Octavia Spencer) notices a new computer has been brought in, possibly to replace her and her staff, and so she immediately went to work to learn how to operate the machine, and taught her staff how to use it as well. Similarly, I am inspired to learn how to use AI tools to enhance my work and increase my productivity.
ReplyDeleteThis module focused on the benefits of AI and how AI is presented in our current zeitgeist. It would benefit from a more thorough review of the major ethical concerns surrounding the foundations upon which these AI systems have been built. This includes stolen and illegally reproduced art and outdated and biased images and image cataloging that once (and still might) form the datasets used to power DALL-E and other AI art systems, the Enron staff emails that were used to build text based AI systems such as Chat GPT, and the fact that so much of our personal data, accumulated through surveillance capitalism, is used to power these systems without the consent to do so.
One surprise for me in this module was in the education I received on how to use prompts to create AI imagery through the Selfie assignment. I found it similar to computer programming – inputting different combinations of text to receive various outputs. By refining my prompts, I realized I was getting better and better at building digital images with the AI tool. It’s not as simple, easy or intuitive as it seems, and the options really are endless.
https://qz.com/work/1546565/the-emails-that-brought-down-enron-still-shape-our-daily-lives
https://www.newyorker.com/culture/infinite-scroll/is-ai-art-stealing-from-artists
https://www.fastcompany.com/90871955/how-generative-ai-is-changing-surveillance-capitalism
Hey Sarah,
ReplyDeleteFurther to your point about having a thorough review regarding AI ethical concerns, it would also be interesting to question what would happen if AI models exhaust all of the internet's available data. Apparently, this is actually happening right now and Big Tech is turning to YouTube content for training data. What could this mean for content creators who don't want their material used to train AI models? Conversely, would some content creators (or institutions disguised as creators) start creating extremely biased content to influence AI? There are so many questions, and I truly wonder if the research and consequential legal processes can keep up with tech's evolution.
Here's the article: https://www.morningbrew.com/daily/stories/2024/04/07/tech-companies-find-the-edge-of-the-internet