What is Generative AI?

Generative AI (GenAI) is simply that: an artificial intelligence that generates new content, from being trained on huge amounts of data (that corresponds to the form the model’s outputs are).

Basically, a model (the artificial intelligence part) “reads” a ton of material that programmers have coded with keywords and other factors to help the model make sense of the content. And it’s a LOT of content, with most models trained on text that would take a human 20,000 years to read[3]. For the models most general consumers are familiar with, once the model has learned the patterns in that data, a user interface is developed that allows you to INPUT a request (a “prompt”) and then the model takes all its knowledge to respond with the best OUTPUT (ie text, an image, etc.).

Full image and accompany description compiled by Arham Islam and can be found here.

Full image and accompany description compiled by Arham Islam and can be found here.

Check out these pages to learn more about the two types of generated content most relevant to this project:

Text Generation

Image Generation

Think this is all new and a bit out of the blue? Not really. GenAI’s development has roots in the Renaissance (when we started to ask how technology can be used to communicate information), Ada Lovelace’s work on the “Analytical Engine” in the 19th century, and computer generated art appeared in the Tate in 1962 [1]. The image to the right covers a bit more of GenAI’s history in depth starting in 1932.

However, while the history of GenAI might be long, we know it’s growing fast, with the expectation that 90% of online content will be synthetically generated by 2026 [2]. GenAI is here to stay, and now is the moment for writers to engage with the tools, learn the benefits and challenges of genAI in their craft, and become an active voice in the conversation around regulation and development. Before legislative bodies and corporations make the decisions for us.

Applications of GenAI Models for Fantasy Writers & Worldbuilders

This notebook will demonstrate the role genAI can play in the development of fantasy worlds through in the following ways:

Key Things for the Fantasy Writer to Understand When Using Text Generation:

  1. Understand a model’s training data
    1. The data that a genAI model learns from is often scraped from the internet (think Reddit, blogs, social media, etc.), which means the model can have all the same “unconscious” bias and stereotypes that humans do, and might have captured copyrighted data as well
  2. AI Hallucinations
    1. AI hallucinations are “false information or untruths with outputs that are incorrect, misleading, or fabricated, rather than being based on accurate or real-world data” [4]
    2. Be clear when you need the model to be accurate (versus just providing a thought starter for you) and fact check everything
  3. Protecting Intellectual Property
    1. Do NOT use the names of artists, writers, etc. or their specific works in prompts if the creator/work is not in the public domain (or you are basically asking the AI to go find and adapt pirated content for you)
    2. Carefully review all terms and service of any tool you are using, so that you understand how your inputs into the tool are being stored/ potentially used to continue training the model

<aside> 🗞️ Favorite Resources to Keep Up with the Latest in GenAi

Google Cloud Tech’s 20 min video on GenAI (great & detailed overview for someone with no AI development experience)

Google Cloud Tech’s 20 min video on GenAI (great & detailed overview for someone with no AI development experience)

Graphic made by George Lawton, original can be found here.

Graphic made by George Lawton, original can be found here.

Citations:

[1] Rupert Breheny, Paul Dowling, Margarita Repina, Samson Vowles, Matt Garbutt, “Generative AI Art- But is it art?” (panel at Gen AI Con, London AI Summit, London, June 14, 2023.)

[2] Ellen Poughia, “Generative AI and Media Publishing; How Gen AI is reshaping content creation, consumption and curation” (speech, Gen AI Con, London AI Summit, London, June 14, 2023).

[3] David Sanchez, interviewed by Killian MacDonald, June 26, 2023.

[4] Patrick Parra Pennefather, Creative Prototyping with Generative AI: Augmenting Creative Workflows with Generative AI. ( Berkeley, CA: Apress, 2023).

<aside> 🧭 Notebook Key 🔊 Listen in for an explanation (audio file) 🤖 Working with GenAI tools 💡 World ideation points 🎙️ Notes from interviews with experts ✍🏼 📹 🎭 Sample snippet of potential content in different forms

Bibliography

</aside>

Following this link to the next recommended page:

<aside> Create Your Own Work

Prompt Library

Tools

</aside>

<aside> Learn about GenAI

Introduction to GenAI

</aside>

<aside> GenAI in Action

Artefact 1: Map & Life in Those Places

Artefact 2: Magical Knitted “Armor”

Artefact 3: Folk Song Turned Musical Theatre Concept

</aside>