This week we focused on creative constraints when working with AI. I found that when I thought about creative constraints, I was thinking more in terms of personal creativity and ethics rather than limitations of software. The discussion in class helped me to realise more constraints than I had initially thought of. Working with AI to create media needs a lot more consideration than simply the prompts we feed into the generative AI. In the past, I would simply list the main features in the imagery/video that I wanted without much thought to other processes such as camera angles, lighting, mood and more. With our upcoming assignment I will be able to think about these conventions more closely to create an even more accurate image/video generation to what I see in my mind. I also learned that Runway actually has specific tools and directions to help guide the AI generation tool which is very useful.

We also had the practice prompt battle in preparation for our next class. I found this activity really enjoyable and I look forward to participating in class again on Monday. Hearing my peers ideas and coming to a unanimous decision can be difficult at times but it is a valuable skill to learn especially in the creative industry which is often collaborative that we are in.

A piece of news I learned about this week, despite it being a bit of old news, was how Apple is integrating ChatGPT into IOS 18 in Siri. I never really use Siri because I found it quite useless but now that ChatGPT will be assisting the software, maybe I will give it a try. In fact I have a Google Home and Amazon Echo in my room but I also find them quite useless except for the Echo tracking my deliveries and the Google Home only for weather and alarms. They were cheap so it isn’t a big loss but if the AI used in these products improves, especially with ChatGPT, I can see AI in my everyday life grow substantially compared to how much I use it now. While I obviously do use AI in my everyday life a lot without realising, I want to see more deliberate choices in my free will to choose to use AI. Until then I feel like AI will just remain an afterthought for me. I had only started to use ChatGPT semi regularly for the first time this year so I am a little late to the trend but I am starting to realise how useful these tools can be.

Another week another blog. This week we took a look at voices in the AI art space for and against the use of AI in art. I am a fairly middle ground person so I can definitely see both sides of the coin. In class there were many great examples of AI art that was found. Théâtre D’opéra Spatial is an image created by Jason Michael Allen using Midjourney and it was the one that stood out the most to me. I had heard of the commotion surrounding the piece at one point but never actually saw the image. It is a very beautiful piece despite being created by AI and I also love the space opera genre. It gave me the vibe of Diva Plavalaguna from The Fifth Element (slightly off topic). Finding this AI generated art found me a new niche I never thought about. I will say that while AI can generate very exceptional art, it does feel as if it’s the exception not the rule and these images are not what we commonly see when AI art is found. I searched for AI artist recommendations in Reddit and all the artists had the very AI feeling art. It makes me question just how was the Théâtre D’opéra Spatial created? Surely it is too complex for it to be completely AI generated.

We have another upcoming assignment next week and I am looking forward to seeing what I can create. Originally, I wanted to create a mini film based on Diva Plavalaguna after seeing the AI art that inspired me but I didn’t want to infringe any copyright. I definitely prefer to create stories and images based in fantasy compared to realism so that is the direction I will head in. Since I had previously done the AI studio Automatic for the People, I have experience creating AI assisted media however we mostly focused on Imagery and moving clips. Although we did use AI generated voice clips, I have never tried to use AI generated sound effects before so this will be a new venture. I prefer not to use AI with my script writing, but I will give it a go and see how it can help guide the process perhaps if I get stuck with my story.

We had four readings this week but I am going to briefly discuss the Panel discussion at the AI Film Festival Los Angeles 2024. One of the major talking points during the panel was about how AI was a positive tool for failure and also how it can allow us to analyse our past work and pivot into new directions. The focus on how AI can be used to drive our creativity and challenge ourselves was a new perspective for me. I only really thought about the ethical implications of AI generated art but there are other ways AI can be used to assist in the creation of art/media without relying on AI generated work.

References:

Panel Discussion | AI Film Festival Los Angeles 2024 | Runway 2024.

This week was a big one for the studio! We had our first assignment due. We had to create the two-minute documentary and do the critical reflection. I feel like with a lot of media I create, especially video content, you can gradually see my effort decrease as it goes on which I need to work on, but I am a master procrastinator and did the assignment the day it was due. I also still find it hard to consolidate information and what are the best points to keep, especially when the max length is short, but it is a good skill to learn as I will sometimes have to face restrictions on word count or video length in the future. Overall, it was a god assignment to share something we use AI for in everyday life and I’m very passionate about video games, so it was nice to create. However, I’m not sure if it was close enough to the outline of a documentary of how I use AI in my everyday life or not.

We then had our in-class presentations. I was not expecting to do such an activity, it was the first time, but it was a good test of how well we work under pressure. I cannot say I work very efficiently with short time limits, and I mostly did my part off the top of my head. Every group had an interesting topic. One focused on how AI is used to detect Parkinsons disease early and help treat it. Others included environmental impact and the use of AI in homes with googles AI. It was interesting to see each groups unique perspectives. While there are some environmental concerns, I believe there is much bigger environmentally impactful issues than AI right now. I do think AI will be very beneficial to us in the future and we are well aware of the benefits especially in the healthcare sector, but I am not really informed on how it could negatively impact us besides environmental impact. I know jobs could be affected and potential security risk but only at the surface level. Maybe AI will gain sentience and destroy humanity terminator style.

The readings for this week mostly focused on how AI will affect creative industries and its implementation into these spaces. The most memorable idea for me was how creatives were reluctantly allowing AI in their workflow as long as it only took over the more mundane and repetitive tasks (Bender 2024). There was also mention of how AI art is not the same as human art and doesn’t offer the same emotion humans can provide through their art. I agree with both sentiments but I do think it’s only inevitable that AI will eventually take on more major roles in not only the Creative industries but other ones too and we will just have to learn to work with it unless there begins to be a big niche for AI-free art.

References:

Bender, S 2024, ‘Generative-AI, the media industries, and the disappearance of human creative labour’, Media Practice and Education, pp. 1–18.

McLoughlin, J 2024, ‘The work of art in the age of artificial intelligibility’, AI & Society.

In week 2 of our studio, we focused mostly on prompting and how we can manipulate generative AI tools to output our desired generations. Using Leonardo.AI we participated in a hackathon with different challenges. I found the first task to be the most difficult, generating a self-portrait. It is pretty hard to get a good representation of ourselves especially when we can only use word prompts and most people haven’t probably tried to describe themselves before. It is a good lesson though in looking at the details and how we can describe things and what is noticeable to us. Although that probably isn’t the aim of the studio, it also helps us to learn how to analyse other things we may want to generate and prompt them efficiently.

I found the AI was refusing to change details about the generated portrait despite me explicitly stating very specific statements such as I have no facial hair or no earrings or even small ears, but the AI refused to change these details. It was very frustrating at some points, but it does make me feel like prompting is similar to coding in a way. When there is a bug, we just can’t figure out, so we have to really look at where we went wrong or try different ways to get the result we need. I did find it better at creating environments and designs compared to real humans. In one image I generated as the poster for the studio, there is a human face behind the Alien/Artificial humanoid and while the figure in front has a normal face, the human face behind it has an absurdly high nose. It just seems to struggle with human faces which is funny since you would think there is so many pictures of people online to feed the AI with.

The reading this week was a French Literature piece which to be honest, I found it incredibly difficult to digest anything. I had to use Chat GPT to simplify and explain many sections of the text which did help. An interesting point from the readi ng was how it’s crucial to understand how consumers interact with products, as individuals in society dictate largely how these tools can be used. This is something that should be taken into consideration with AI tools as the way society and different cultures repurpose the tool will change how it’s used.

I had previously taken part of an AI based studio last semester being Automatic for the people and comparing it to this studio, I found this studio has a more in-depth dive into what exactly generative AI is and what we can do to use it effectively. The coding aspect in the first week was a great way to show us a glimpse into how a generative AI program can be developed.

Here is some of the images generated.

References:

de Certeau, M 2011, ‘III. “Making Do”: Uses and Tactics’, in The Practice of Everyday Life, University of California Press, Berkeley.