Stopping the Birthday Project

I am considering ending the AI-assisted Birthday Project, and I need to figure out if I should or not. There are five issues which make me a little uncomfortable.

First, this:
https://www.washingtonpost.com/technology/2025/06/11/disney-universal-sue-midjourney-ai-copyright/
"Disney and Universal’s lawsuit frames the issue as a matter of good versus evil, calling Midjourney “a bottomless pit of plagiarism.” AI industry advocates counter that legacy media companies are standing in the way of a technological advance that could unleash a wave of creativity."

1) I am not bored by it
I have loved doing this for my family. There is one member left of my parent's generation, there are 8 of us in my generation, there are 8 in what would be my child's generation if I had any of those, and there are 5 in what would be my grandchild's generation (22 including spousal units), and it has been delightful getting know each person and keeping up with their processes, and I'm not going to change that going forward. It seems like it was a good thing — everybody likes to be recognized for who they are, and for me, finding out who they are and designing a visual joke for them each year remains one of the most creative tasks I have ever performed.

2) I am not afraid of "stealing from other artists" (a big one)
Using AI, I have been directly accused of the theft of other people's artistry, in particular styles and concepts, theft which is sometimes but not always recognized or reported by the AI project. Well, that's what art is all about, and I would argue it is not theft. It's art. I've been listening to Dolores O'Riordan and the Cranberries lately, and her musical style reminds me of a whole range of others — Alanis Morissette and Scandinavian singers like Aurora. The reason they sound alike, I'm told by my husband, is because they use Irish keening, an ancient collection of motifs and song styles. Every artist on the planet learned their trade by copying their forebears, and leaping off of the shoulders of giants.

I've also been accused of putting artists out of business, but that's clearly bogus, too. I'm never going to be in a place to pay an artist. I give away my pictures under a 4.0 Creative Commons licence. I'm clear where I get my images, I credit where the style comes from and I avoid using living artists or an artists still under copyright. If you ask me how would I feel if my artistic creations were stolen by AI, well, I already suffered that. The early AI programs explicitly listed ThoughtCo as a place where they were training their AI models. I lost my job writing for ThoughtCo after 20 some years, primarily because of budget cuts, probably set forward because of cheaper producers. I found another venue and kept writing until my brain got fuzzy. As a creative, you know that having a paying job is a blessing but not a promise that can be kept in perpetuity. There are always costs to you when you're paid — people who pay you get to tell you what and what not to write/draw/create.

However, mostly to entertain my sister's grandchildren (not to mention the rest of us), I've used ideas, and some image clonish ness from modern pop culture, including Star Wars, Minecraft, Zelda, Frozen, Nightmare at Christmas, Paw Patrol, and Harry Potter. So, as Jerry Lundergaard said (William H. Macy in Fargo), there's that.

3) the personal cost is not totally prohibitive
I am on the "standard" plan of MidJourney — $305 / year, and now that I see that I think that is too much. I have access to Google Gemini, which switch would allow me to spend that money to do something charitable.

4) the effect of the process of making the images on the environment is worrying
One truly disturbing issue is the environmental costs of Large Language Models. The main culprits are CO2 emissions, water consumption, and e-waste, which have been found during the training process but also during production. Because these issues directly impact the producers, researchers are currently working on making the process more budget and environmentally friendly.

• Li, Zihao et al. "The Digital Revolution's Environmental Paradox: Exploring the Synergistic Effects of Pollution and Carbon Reduction Via Industrial Metamorphosis and Displacement." Technological Forecasting and Social Change, vol. 206, 2024, p. 123528, doi: https://doi.org/10.1016/j.techfore.2024.123528
• Parsons, Jacob. "Llms and the Effect on the Environment." EVIDEN News and Insights, 2024. https://eviden.com/insights/blogs/llms-and-the-effect-on-the-environment/
• Shu, Lee-Lean. "The Untold Story of AI's Huge Carbon Footprint." Forbes, 2024, https://www.forbes.com/councils/forbestechcouncil/2024/04/26/the-untold-story-of-ais-huge-carbon-footprint/
• Team, Holistic AI. "Ai and Esg: Understanding the Environmental Impact of Ai and Llms." Holistic AI, 2024. https://www.holisticai.com/blog/environmental-impact-ai-llms
• Wang, Weilong et al. "The Cost of Pollution in the Digital Era: Impediments of Air Pollution on Enterprise Digital Transformation." Energy Economics, vol. 134, 2024, p. 107575, doi https://doi.org/10.1016/j.eneco.2024.107575
• Yuan, Honglin et al. "The Impact of Digital Economy on Environmental Pollution: Evidence from 267 Cities in China." PLoS ONE, vol. 19, no. 1, 2024, p. e0297009, doi https://10.1371/journal.pone.0297009

5) the reader can't always tell when an image is faked
Another worrying thing: the proliferation of fake images (possibly, likely, may, might, pick an adverb) make people doubt the reality of photography in general. If there is a video of a particularly high-status individual doing or saying something treasonous, there isn't currently a way to verify if the file is real or faked. This is a critical issue, but computer scientists are currently working on ways to digitally watermark files (audio and visual) with information about how and when the data were created, and can persist against attacks to remove that data.

• Amrit, Preetam and Amit Kumar Singh. "Survey on Watermarking Methods in the Artificial Intelligence Domain and Beyond." Computer Communications, vol. 188, 2022, pp. 52-65, doi: https://doi.org/10.1016/j.comcom.2022.02.023
• Lancaster, Thomas. "Artificial Intelligence, Text Generation Tools and ChatGPT – Does Digital Watermarking Offer a Solution?" International Journal for Educational Integrity, vol. 19, no. 1, 2023, p. 10, doi: https://doi.org/10.1007/s40979-023-00131-6
• Regazzoni, Francesco et al. "Protecting Artificial Intelligence Ips: A Survey of Watermarking and Fingerprinting for Machine Learning." CAAI Transactions on Intelligence Technology, vol. 6, no. 2, 2021, pp. 180-91, doi: https://doi.org/10.1049/cit2.12029

So: the bottom line is, there are serious concerns about these kinds of images, but there is little I can do personally about them, and luckily, any issues that I can think of other people have already addressed and are already addressing. I just have to decide if I am at a point where I can actually trust that the issues will be resolved eventually.

Options:
Continue as before with Midjourney
Get out of Midjourney and continue using Gemini (cheaper); timing is bad, because MidJourney could probably stand to hang on to my business.
Tailor my images more carefully to not impede upon big shot companies. Really?
Continue privately but post nowhere public. (I can hide Wasteflake images).
Stop. Just stop.

Note: Judy Holliday makes me happy. From https://aurorasginjoint.com/2024/05/24/introducing-jack-lemmon-a-guy-youre-gonna-like/