By Mya Keay
In this Module we were instructed to look into and explore the world of generative AI. The use of AI over the past two years has increased dramatically since the introduction of Chat GPT on November 30th, of 2022. Since then new technologies and AI platforms have been popping up everywhere. As we are meant to learn about generative AI systems that are new to us, I decided to use the website Craiyon.com for the assignment.
What is generative AI?:
Before delving into my experience using Craiyon, I wanted to first establish a base point of what Generative AI is. Generative Artificial Intelligence is a system or type of artificial intelligence that aids in creating new ideas, art, music, presentations, or simply just assists the user with tasks such as grammar or video editing. The main difference compared to general or traditional AI is that generative AI is automated. Meaning it can create something that hasn’t been started or made yet. You could ask generative AI to create a presentation or image with very few prompts and it will automatically generate what you are asking for. Be it a song, a short story, an essay, a presentation, even a piece of art.
My use of generative AI:
The generative AI I chose to use for this assignment was Craiyon.com. Crayon.com allows users to type in a prompt and within 60 seconds be given an array of artwork generated by AI to match the prompt given
The image above was the first attempt I had using this website. I started off with a basic prompt “draw a pig playing music”. What I realised from the image above was that being as my prompt was simple I would get a simple image in return. I had not specified how I want the pig to look or what instrument, therefore the pig is holding a string, the music notes are more so squiggles and the face is slightly deformed. I decided to ask again but this time giving more detail, asking for “A joyous pig with a big smile playing the drums.”
As shown the more detail I gave the better the quality I received. What I learned from this is that when using AI you need to be specific in what you are searching. AI is unable to distinguish real from fake so in order to get the results you are wanting you need to specify what it is you want.
What would I use AI tools for in the future and what would I not use them for:
In the future I would definitely use AI tools such as ChatGPT, Crayon, and others to help create a general idea of a project perhaps. If I was tasked with drawing a dog wearing sunglasses for an example, I would ask Generative AI to draw me an example, not to use, but to have as a guide. The same goes for tools such as ChatGPT. I would use it to help me brainstorm but not to write a portion of the assignment. That being because Generative AI and AI as a whole is not always 100% factually correct. Programs such as chatGPT only have the information that they are programmed with. If a question is presented that the system does not know the answer to 10-20% of the time it will make up an answer based on what it predicts the answer could be based on the information it does know. Making it not truly a reliable source. Therefore I would use AI to aid in the creation of Ideas and brainstorming but would not rely on it nor use it to create an assignment or presentation that I would then show or submit.
That being said, I have used Chat GPT to create a SMAR analysis on the Craiyon website.
SMAR analysis on Craiyon.com:
Before using AI to create an analysis I created my own breakdown to see if what I received was accurate to what I had believed about the site and how it fit into the SMAR model. I believe that the LLM (Large Language Models) was able to accurately capture what I had thought about the site.
The SMAR model helps users to understand how the technology being used can help enhance learning and creativity.
- Substitution:
Craiyon uses a text to image generator. This allows users to provide prompts to create an image, and then be given art to match the prompts provided. This substitutes the need for the physical side of producing art, such as manually drawing.
- Augmentation:
Instead of taking the time to produce a prototype by sketching which they would then alter and refine over time. The site uses augmentation to provide instant results which users can then modify by increasing or correlating the specifications.
- Modification:
Because users provide the prompts they can alter or change their prompts as often as they’d like to explore other pathways or options. This allows for an endless amount of modifications in order to find what sticks with you the best.
- Redefinition:
It opens the door for users to express their creativity in a way perhaps almost impossible by hand. As users can generate extremely imaginative and detailed images, they are able to explore ideas that might not be achievable otherwise. It also allows people to engage in abstract ideas which can make learning with this tool more interactive and engaging.
Craiyon demonstrates how generative AI can not only replace traditionally creative methods but also enhance people’s creativity by allowing them to use new abstract forms of expression.
“How does the Craiyon website use substitution,augmentation, modification, and redefinition” prompt. ChatGPT4.0, 14 Mar. version, OpenAI, 10 Oct. 2024, chat.openai.com/chat.
What are the strengths and weaknesses of the site?:
Strengths
- User friendly: Crayon has a simple design that makes it easy for all users to navigate.
- Free to use: it allows users to generate endless images for free.
- Ability to share: the site allows users to share the creations made to others, allowing people to engage with others
Weaknesses
- Image quality: the quality of the image is less sophisticated and images are not always accurate.
- Site limitation: because the site is free it has a much longer leading time to receive images unless you pay for an upgrade.
- Limited editing: although you can alter your prompts you aren’t able to edit the images as a whole
- Theft: because it uses images from around the web and combines them there is risk of copyright issues.
Recent Comments