Blog post 1: https://www.mediafactory.org.au/chloe-stelling/2024/10/23/assessment-4-blog-post-1/
Blog post 2: https://www.mediafactory.org.au/chloe-stelling/2024/10/23/assessment-4-blog-post-2/
Blog post 3: https://www.mediafactory.org.au/chloe-stelling/2024/10/23/assessment-4-blog-post-3/
Blog post 4 and final video: https://www.mediafactory.org.au/chloe-stelling/2024/10/23/assessment-4-blog-post-4-and-final-video/
Link to video: https://vimeo.com/1020939874?share=copy
Link to production dossier: Chloe Stelling s4001537
My video explainer revolves around the question ‘how is the rise of AI (artificial intelligence) and ADM (automated decision-making) impacting creative writing industries?’. Throughout creating this video, I’ve come to realise that there is no clear answer to this; the situation regarding how AI is now becoming a recurring part of all industries is not black and white. There are pros and cons to having this technology infused in the creative world. For example, a big concern is that major Hollywood studios have reduced the hours and wages of screenwriters, instead favouring AI generated material. This significantly impacts the livelihoods of writers, violating their occupation and creative freedom. An advantage however, is that it can help creatives enhance their work, reducing editing and other tedious tasks. My video aims to explore what the increase of AI and ADM in creative writing will do to the sector in the future.
My process started with a lot of research about the Writers’ Guild of America (WGA) strike in 2023, when thousands of Hollywood screen writers went on strike for many reasons, one of which was to protest against the major studios’ favouring AI generated written content instead of human made work, which reduced their wages. This was a starting off point for me to research more about how AI and ADM are affecting the creative writing industries, which is particularly personal for me as an avid create writer myself. I investigated the negatives of the situation but also the positives, as AI and ADM can provide some exciting opportunities for professional creatives. I also found that some parts of my script and research were informed by my interviewee Meg, who gave interesting insights into the topic, and into other creative industries as well, that I hadn’t considered. As for filming, I did stick to my shot list and script though I used more found footage than I planned. This isn’t necessarily a bad thing, because it was all relevant and served a purpose. If I was better at managing my time I would have filmed more b-roll myself.
Work in Progress – For making
Regarding making the video, time has been the biggest challenge I’ve faced. During this busy semester, I’ve struggled with time management and unfortunately the filming may have suffered from this. I really wanted to include handwritten text in my video to have a homemade and artsy aesthetic, but with time running out I’ve had to eliminate this idea. Which is a big shame for me because I was looking forward to it. I have conducted my interview with Meg Herrmann, a PhD candidate at university of Queensland researching how AI and ADM impact aesthetics in film and television. The interview was very successful! Meg had a lot of insightful things to say about how AI and ADM affect creative industries, particularly about ethical considerations and how we can smoothly integrate AI into human work and environments. The interview went for 40 minutes, so I had to examine which parts were the most useful and extract them.
When showing Ruth my rough cut, I hadn’t edited it together yet and hadn’t filmed my piece-to-camera, so I showed her my interview with Meg and she helped me pick out sections of it that were particularly interesting to put in my video. I enjoyed showing my footage to someone else to get her perspective, though I was stressed because I still had quite a lot left to film. Most of the class actually couldn’t make it to the rough-cut screenings so it was disappointing that I couldn’t see the work other people had done. Nevertheless, I’m grateful that Ruth could provide me with feedback and guidance on how to proceed in the next steps. With some of the found footage I’d collected that I plan to use as my introduction (overlapping news reports about AI and the writers’ strike), Ruth suggested that I trim these down at the beginning and end to make them sharper instead of playing the whole clip, creating a punchy opening that immediately hooks the audience. I wish I had more footage to show her but like I mentioned, I’ve been struggling with time management.
Work in Progress – For researching
In terms of research, I have included a range of journalistic writing, personal opinion pieces, and other video explainers. I’ve had to do lots of research about the WGA strike in 2023 as it is a large focus in my video. There was so much information about this strike that it was somewhat overwhelming and a challenge to try and condense it succinctly. Very interestingly, I actually used AI to help me summarise my research, which is fascinating because my whole video explores the idea of how AI is impacting the writing industries. Moreover, I found some opinion pieces written by Linda Maye Adams to be particularly insightful. She strongly condemns AI in creative writing when it is used to simply generate stories. She tested AI programs in creating stories and assessed them to be “bland”, and had merely “mined the internet” to find sentences. This idea of testing the AI programs to see how it compares to human writing is something I’d like to include in my video. I did experiment with this idea and ran some prompts through ChatGPT such as, give me an introduction to a coming-of-age story, and give me an introduction to a psychological thriller story. Something extremely fascinating I found through this experiment was that ChatGPT included some very similar, near identical descriptions despite being two completely different narratives. I found that incredibly interesting, and perhaps I can investigate this phenomenon further. I think that experimenting with AI programs counts as research, because I spent a lot of time running prompts by ChatGPT, asking it for creative writing starters, character descriptions, dialogue, etc, and used this to draw some conclusions which I will include in the video. I found that AI generated creative writing is very cliché and includes a lot of over-used descriptions and dialogue pieces, most likely because it has just scanned the internet for ideas and plucked out the most frequently occurring ones. This goes back to the idea that both Linda Maye Adams and Meg, my interviewee, have stated; that AI cannot actually create or come up with original ideas because all it can do is access the already existing data it is trained on.
References
Adams L M (2024) ‘The Problem with AI and Fiction Writing’, Linda Maye Adams: Speculative Fiction Author, accessed 26 September 2024. https://lindamayeadams.com/2024/05/05/the-problem-with-ai-and-fiction-writing/
Adams L M (2023) ‘Artificial Intelligence and the Future of Fiction Writing’, Linda Maye Adams: Speculative Fiction Author, accessed 26 September 2024. https://lindamayeadams.com/2023/12/17/artificial-intelligence-and-the-future-of-fiction-writing/
Los Angeles Times staff (2023), Writers’ strike: What happened, how it ended and its impact on Hollywood, Los Angeles Times website, accessed 24 September 2024. https://www.latimes.com/entertainment-arts/business/story/2023-05-01/writers-strike-what-to-know-wga-guild-hollywood-productions
OpenAI (2024) ChatGPT, [Large language model], accessed on 30 September 2024. https://chat.openai.com/chat
Reflection on week 9 – Work in Progress presentation
The main feedback I received from our guest panel member Steph during the WIP presentation was to focus my video explainer on how AI and ADM are impacting creative writing industries instead of broadly speaking about all creative industries. She felt that I was more passionate about the writing aspects and would have the opportunity to explore this in more detail if I centre the video around it. I agree with this and had felt myself leaning towards this idea as well. Steph gave me two great ideas that I am excited to incorporate. She suggested that I could have AI write some of my interview questions and see how effective they are compared to the ones I write myself, which poses the question to the audience: can they tell the difference between human and AI writing? She also suggested that I use the Writers’ Guild of America (WGA) strike as a hook into the video to intrigue the audience before revealing the actual topic. When preparing and presenting my research and ideas, I felt that I had a lot to cover and not a clear way of linking it all together, so the feedback really assisted me in narrowing down my topic to make the ideas flow cohesively. Whilst it was daunting to share my work in progress with the class when I didn’t feel entirely confident about my topic, I benefited from hearing the constructive feedback which I feel will strengthen my video. Furthermore, I found Marlow’s presentation to be very interesting; his video will revolve around the idea of Melbourne as a smart city and the Internet of Things (IoT), and how data is being collected from us literally everywhere we go. I really enjoyed his presentation and am intrigued by his idea as it is something I’ve never really considered. His research stated that data is being collected from our phones, watches, laptops, etc, every time we walk somewhere or how long we stay somewhere with Wifi, which tracks our movements and is used to train other AI models. I like his idea of filming a walk through the city that will show what aspects of it are ‘smart’ and how they work. His approach to choosing a topic is different from mine and I like how relevant it is in today’s society.
Blog post 1: https://www.mediafactory.org.au/chloe-stelling/2024/09/01/decoding-ai-assessment-2-post-1/
Blog post 2: https://www.mediafactory.org.au/chloe-stelling/2024/09/01/decoding-ai-assessment-2-post-2/
Blog post 3: https://www.mediafactory.org.au/chloe-stelling/2024/09/01/decoding-ai-assessment-2-post-3/
Video and reflection: https://www.mediafactory.org.au/chloe-stelling/2024/09/01/decoding-ai-assessment-2-video-and-reflection/
Our small video explainer examines AI in animation, including how AI animation generators get their data. Overall, the process of making this was insightful and interesting; animation is not something I have much (or any) experience with so it was fun to research and learn about AI in animation. I think our topic worked well with the requirements of the task given that it was supposed to be a simple video explainer. Our topic was quite basic, but also has room for expanding it in the future as there is so much more information we could cover. In terms of collaboration, I believe our group was compatible and worked well together. We have similar interests regarding art/media and AI. I do believe, however, that we lacked effective communication and struggled to work efficiently on this project outside class time. That being said, there were also external factors that influenced this, such as illnesses and other assessments being due. Perhaps in the future we can improve on time management. Overall, though, the collaboration experience was a positive one. Furthermore, if we were to develop this piece in the future, I would love to make a proper example of an AI generated animation to show the audience. We made small AI generated animations that were basically still images, so I think it would be really interesting if we made a longer animation with characters and dialogue to see what AI would produce.
Link to video: Decoding AI final video assessment 2.mp4
The third blog post should reflect on a reading and discussion topic from class from weeks 4 – 6 (you may choose one that you feel most drawn to). What were some insights gained from this reading? Draw connections between the reading, discussion and your own experiences/insights from the everyday.
In week 5 we explored the idea of AI and automation in relation to health. Sadowski (2024) examines AI health coach apps, including Thrive AI Health, which supposedly provides users with personalised advice and recommendations on how to make one’s lifestyle more healthy. Sadowski is sceptical of AI health coaches and does not believe in their ability to “solve the world’s chronic disease problems”. Firstly, he proposes that AI health coaches are riddled with the same biases and errors that are common in other AI models (as well as human doctors and institutions), therefore their effectiveness cannot be guaranteed, despite the thorough personalisation. Furthermore, he asserts a major flaw in the idea that AI health coaches can cure widespread illnesses in society – which is that all individuals choose to make certain decisions without external factors influencing or forcing their actions. The AI health coach logic relies on the idea that individuals are choosing to engage in unhealthy eating habits which can be changed by a simple reminder from an app. In reality, many people cannot afford or do not have access to fresh, nutritious food and consume unhealthy foods because they are cheaper and more accessible to healthier options. In these instances, having an AI health coach telling someone to eat healthier will not actually make any difference. The “social determinants of health” influence a person’s quality of life and health, including their access to health care, fresh and nutritious food, ability to exercise, work conditions, etc. Sadowski criticises AI health coaches for not taking the social determinants of health into consideration and merely seeing the surface of a person’s health choices, which they may not always have a lot of control over. I found this idea very insightful as I had not previously considered this. In this instance, AI is not particularly helpful; it can provide recommendations and reminders but this does not change the social and economic factors that can influence a person’s health.
Sadowski J (12 July 2024) ‘Why an ‘AI health coach’ won’t solve the world’s chronic disease problems’, The Conversation, accessed 16 July 2024, https://theconversation.com/why-an-ai-health-coach-wont-solve-the-worlds-chronic-disease-problems-234369
The second post should reflect on your current research/work in progress for your individual video project. What are some potential ideas/topics you’d like to cover in your video explainer (e.g., if your video is about recommender algorithms, do you have a case study or example you want to include?) Make connections between your research and things you have encountered and observed in your own media engagement (e.g., have you read some research about social media algorithms that you can connect to your own social media use?)
At the moment, the topic for my video explainer revolves around art and AI, more specifically, what does the rise of AI and automation mean for creative industries? I had originally planned to focus the explainer video more so on film and television, but in my video explainer proposal, I discussed a lot about graphic design and how companies may see AI as a cheaper, quicker, and more efficient option when it comes to any artistic content a business might need. I still have some more research to do, as I would like to still include discussions about AI in film and television as well as design and physical art. Some case studies include the use of AI in script writing for television. I would like to discuss the Writers Guild of America Strike in 2023, when script writers went on strike to protest the use of A in screenwriting. Writers had major concerns that AI was being used to replace them and protested against studios using AI instead of them, which would cut their wages and violate their creative rights, especially if the studios decide to use AI to edit or change work written by a human. After researching this strike, I’m very keen to discuss more about it in the explainer video, as it was such a significant development in the rise of AI and automation for the creative industries. The writers won the battle and were given new terms that would protect their creative rights and wages against AI services that threatened them. An economist at MIT, Simon Johnson, states the new terms are a “fantastic win for writers”. A key takeaway I learned after researching the terms was that AI is under the control of writers to use it how they please, including to help them write or edit scripts. AI is not under the control of the studios who could take advantage of it to reduce the wages and creativity of writers.
Anguiano D and Beckett L (2023) How Hollywood writers triumphed over AI – and why it matters, The Guardian, accessed on 29 August 2024. https://www.theguardian.com/culture/2023/oct/01/hollywood-writers-strike-artificial-intelligence
The first post should reflect on the ‘genre’ and form of video explainers. How would you define a video explainer (can we call it a ‘genre’, or does it fit neatly into a ‘genre’)? From the examples we’ve watched in class, what are some of the aesthetics, modes or conventions that stand out to you? Choose an example (from class or your own research) and critically discuss those elements that you think work well/what you would do differently? What elements might you like to include in your individual video explainer (based on your assignment one proposal).
We could potentially call video explainers a genre, as they contain consistent elements despite differences in topics. These include lots of voice-overs, piece-to-cameras, interviews, animations, graphics, and acting. There are also genres and styles within video explainers. Some are more analytical and serious, with a target audience at adults or students. These generally don’t include too much acting or storylines and tend to have a more academic tone. These can include interviews with experts or people with an experience with the topic. Other video explainers are light-hearted or funny. These tend to be directed towards children but not always; some older people may prefer to learn through a more casual approach. These can include the use of acting, sometimes with comedic undertones to keep the audience engaged. They can also incorporate easy-to-follow animations, with characters or stories to help explain the topic. For example, we had to share a video explainer with the class, and I chose one created for children by NASA explaining what black holes in space are. This video utilised animations and graphics, most likely because real-life footage of NASA content probably would not interest children, and the animations are a better way to explain a complex idea. Moreover, real-life footage of black holes may not actually show what they are very well, or may not even be accessible. In this case, animations and simple graphics are a better way to teach kids this topic (and older people as well, as it simplified a complicated topic that may be difficult for anyone to understand). If I were to make changes to this video, I would probably speed it up a bit. The voice-over was a bit slow, which is understandable given that it was made generally for children, but from the perspective of a university student I feel like the pace of the speech could be quicker. Sometimes I found myself losing interest because it dragged on a little bit. Overall however, I found it was a helpful resource for learning the basics about black holes in space. I could potentially like to include animations and graphics in my own video explainer, because I believe it was a useful visual tool to help explain a topic. I also will most likely include voice-over because it is an efficient way to convey information with visuals on the screen.
NASA Space Place (20 December 2013) ‘What is a Black Hole?’ , NASA Space Place, accessed on 23 August 2024. https://www.youtube.com/watch?v=OfMExgr_vzY&list=PL9TFrgFq75565gdN1T95J91ciMk7szjiZ