When 1000 designers prompt for "Weather App", what do you think, how variable or how similar will the results be?
Welcome to this edition of the SUX Newsletter. This month I am reflecting about Creativity in the age of Gen AI, a topic that needs to be discussed much more in my opinion - and the introduction of ChatGPT 4o has shown that - again. And we have a bunch of new resources, a list of upcoming events, a new podcast episode and two new Masterclasses for you. So, what can you expect in this newsletter:
- My Thought of the Month - this time about Creativity in the age of Gen AI
- New Resources and and an update about the SUX Playbook
- Latest episode of the SUX Podcast with Gerry McGovern
- Two new dates for our Masterclass "Become a sustainable UX Designer"
- Upcoming SUX Events or conference appearances 📆
- And as always we round up with a little “Did you know” - this time about all the Ghibli-style images that have been created in the past weeks 💡
🌿 Thought of the Month: Future of Creativity
What is Creativity in the age of Gen AI? Big tech wants us to believe that everybody now can be super creative. But, is that really the case? Does it make me creative if I can create an image with one prompt? I can understand that it might feel creative. And that it might feel enabling, especially if I do not have the skill to create such an image without Gen AI. But the question stands: how creative is it?
I want to look at those questions from two different angles today.
The use of Intellectual Property
A Gen AI model cannot create output out of nothing. To be able to respond to my prompt with output it needed to be trained with tons of data. In case of image creation, millions of images were used for the training of the AI - fed into the machine. Without this training the model would not be able to create any output. The problem is: much of the data, of the images, the texts, the books, etc. that is used for training is intellectual property of someone else. We have seen millions of Ghibli-Style images in our social media channels in the past two weeks since the release of ChatGPT 4o. This was only possible because original images have been used to train the model. Without the original work of Studio Ghibli, this would not have been possible. But, Studio Ghibli did not get one cent for their work being used for training, nor have they been asked for permission.
The value of ChatGPT (and all the other big public Gen AI models) is based on intellectual property of real designers, artists, writers, photographers, without paying for it or even asking for permission. And this is not an unintended mistake. This is common ground in the training of the big Gen AI models. Meta used the biggest pirate library LibGen for the training of their Llama AI. This library includes 7,5 million books and 80 million research papers. Meta used it on purpose to avoid negotiations for the licensing and to avoid creating a precedent. And side fact: it is highly expected that LibGen has been used for the training of all other big Gen AI models.
So, the question is how creative is the result of my prompt, if it is based on stealing intellectual property of other artists, designers, writers, photographers?
When Figma introduced their new AI feature last summer, it turned out that it has been training with the App-Designs of popular apps So, when you asked it to create a "weather app" you got something that looked very very very similar to the Apple weather app. And when you made the update by default you agreed that Figma may use your design for training their AI. You actively had to opt out of this - what many did not even know.
So, it's not only the other artists, designers, writers, photographers - it's also us.
The uniqueness of the "Creation"
And that brings me to another point. How unique is my "creating"? Am I creative when I create a Ghibli-like image like millions of other users? It might feel creative, but in my opinion it is not. And it gets worse: because here it is still a human decision to create similar things. People actively are following a trend. And let's be honest, that is so human ☺️
But, what if we do not follow a trend and still the output is not as unique as we might think? As described above, Gen AI models are trained with tons of data. During the training process the model "learns" the connections between data points from these datasets and it learns which data points are more often in the dataset than others. The more connections a datapoint has and the more often it is in the dataset the more important it gets and the more likely it will be part of a future output. It is a strong signal. In difference to a weak signal (Disclaimer: this is a very very short and simplified description of how this works). The point is, when we prompt a Gen AI it always tries to return the most likely result. So, the strong signals. I will not, by design, return the weak signals if we do not very explicitly ask for it. As a friend of mine once said: Strong signals prevail, while weak signals vanish.
In my Figma example from above when asked for a "Weather App" the strongest signal was the Apple weather app (this is again a simplified description) and therefore was returned.
My point here is not the intellectual property problem, but the uniqueness of the result. When 1000 designers prompt for "Weather App", what do you think, how variable or how similar will the results be?
With every prompt we need to ask ourselves how unique, how creative the result really is or how many other people might have gotten something very similar to my "creation".
Creativity and AI
So, is working with AI not creative in general? I personally think we can be creative with AI. But it's never one prompt. It's the process. If you ask artists who create with AI you hear a similar story very often: Their process still takes as long as without AI. It's using it as an enabling tool in my process instead of the magic wand that does everything with one prompt. And we must be aware that the output - by default - is a lot less creative than we assume.
The price of using AI
But before we start using Gen AI in our processes we should ask ourselves something else. What is the price and who needs to pay for it. And I am leaving out the outrageous environmental price here (that is an article on its own, maybe I should start a series..), but only talk about the problem of intellectual property. The rise of the big public Gen AI models is based on stealing intellectual property of others. And I think we should start thinking much more about which models we use and which not.
Because, if I know how OpenAI, how Meta, how Google trains their models (on purpose) and still use it, isn't that also a statement? A statement that I accept these practices.
It's not AI taking our jobs. It's the way Big Tech is treating our intellectual property. And in the end it's ourselves, if we continue using their models uncritically or even one step further, if we continue using them at all.
📚 SUX Playbook and SUX Resource Database
The SUX Playbook is the making. I am working on that at the moment and hope to release a first version for free soon. Meanwhile we have a bunch of new resources on the resource database. Special thanks to SUX community members Oliver Jahn Marina Gröpel (noizybutpurple) and Tadeas Adamjak for submitting new resources. 💚
Check out our notion space to have a look at all the latest additions to the database: https://suxnetwork.notion.site/
But, we will also continue to add all kinds of resources and tools from other people to the database and today I have three new recommendations for you:
📚 Book: Cancel Culture in Climate by Jenny Morgan
This is a great book by Jenny Morgan. I came across it by a posting of Katie Patrick 🙏 I find in super interesting because conversation and dialogue are fundamental for driving change. Finding common ground, staying in conversation is crucial. And though we cannot discuss with everybody we must be careful not to cancel everybody who is not 100% in line with us.
📓 Article: What is SUX and why we should care
As I am working on the SUX Playbook at the moment, I am releasing parts of it step by step on the SUX Website and in our Notion space over the next weeks. You can already find some of the tools of the Playbook in our Notion Space (here) like "The Principles of Sustainable UX" or tools like the "Needs to Consequences Mapping", "Sustainable User Journey Mapping" and more.
🏷️ Cards: Life-centred Design Innovation Cards by Damien Lutz
Check out Damiens latest addition to his great Life-Centred Design Lab. And I want to use this to announce that Damien will come to Europe in June and that we will setup Meetup and maybe in addition a workshop together in Hamburg. Stay tuned, details will follow soon.
🎙 Latest episode of the SUX Podcast: Gerry McGovern
For the latest episode of the SUX Podcast I was joined by Gerry McGovern. Gerry is besides many things the author of the book World Wide Waste and well known for his work regarding digital sustainability and especially digital waste and e-waste. For me this was also really special because I relate to Gerry and his work pretty often in my talks. “Up to 90% of all digital data is not used” is one of the famous quotes from his book World Wide Waste and you probably have heard it somewhere already. Gerry and me had a really deep conversation about the dark side of tech, but he also made crystal clear that it is worth fighting and that he thinks it’s worth to work on changing the world, because as he said: „Whatever I say does not mean I am giving up.“ I can deep from my heart recommend this episode of the podcast.
SUX Podcast Episode 18: World Wide Waste with Gerry McGovern
My next guest will be Maria Guidice the author of "Changemakers". The episode is already recorded and will be released within the next days.
🎓 Two new dates for the Masterclass "Become a sustainable UX Designer" in April and May
For April and May we set up two new open (and updated) Masterclasses. As always we offer a limited number of seats for a significantly lower price for people in low-income situations or from countries with lower incomes. This works on a trust basis, so please decide on your own if you are eligible for these tickets.
What’s new: while the Masterclass takes 4 hours, we offer now an extra 30minutes at the end to have more time for questions and conversation with Thorsten. So overall the Masterclass now takes 4,5 hours.
Option 1: 24th April 2025, 5:30pm-10:00pm CEST, Online, Tickets and all Infos
Option 2: 7th Mai 2025, 5:30pm-10:00pm CEST, Online, Tickets and all Infos
Team-Masterclass for Design- or Product-Teams: https://sustainableuxnetwork.com/academy/masterclass-sustainable-ux
📆 All upcoming Events:
Find below a list of upcoming local and online events of the SUX Network as well as conferences we visit. If you are at one of the conferences, drop me a message and let's have a coffee together.
April:
- Copenhagen: SUX Dinner by SUX Copenhagen, April 28, more infos soon
- Online: SUX Masterclass “Become a Sustainable UX Designer", April 24, All Infos
May:
- Online: SUX Masterclass “Become a Sustainable UX Designer", May 7, All Infos
- Online: SUX Monthly, May 8, invite will be shared in the SUX Slack
- Munich: DDX Conference, Panel about Sustainable UX and Design, May 10, Munich, All Infos
- Paris: FLUPA UX Days, Keynote "Each Design is a Manifesto for the Future, May 23, All Infos
June:
- Hamburg: SUX Hamburg joined event with the Design Zentrum Hamburg as part of the Sustainability Week, first week of June, more infos soon
- Aarhus: UX Connect, Opening Panel, June 10+11, All Infos
- Hamburg: SUX Hamburg Meetup and Workshop together with Damien Lutz, June 18 or 19, more infos info
July:
- Munich: Green IO Munich, Talk + you can meet us the whole day at our SUX Booth, July 3, All Infos
🌿 SUX Keynotes, Workshops or Trainings for your organization:
Are you looking for a keynote, a workshop or a training session about Sustainable UX or Responsible AI? We offer these for any kind of organization. In fact this helps to finance the SUX Network and all of its non-profit work. If you want to learn more, please get in touch: info@sustainableuxnetwork.com
💡 Did you know:
A Topview analysis reveals that generating just 1 million Ghibli-style images consumes approximately 40,000 liters of water, a volume that, according to the United Nations’ minimum water requirement of 20 liters per person per day, could sustain 2,000 people for a day.
Source: https://www.onfocus.news/1m-ai-images-water-for-2000-people-still-worth-the-likes/
Thank you for reading all the way down here. 🙌
If you have any feedback, thoughts, want to share something, please reach out to me on Slack or on LinkedIn. 🙏
See you,
👋 Thorsten
Founder of the SUX Network & Co-Author of the Web Sustainability Guidelines, Keynote Speaker, Sustainable UX Trainer, Digital Sustainability Trailblazer and Responsible AI Advocate. Get in touch if you want to work with me or book me for a talk, workshop or training. And follow me on LinkedIn, where I share frequently tipps, resources and thoughts about Digital Sustainability, Responsible AI and Sustainable UX.
🌏 SUX Community Count:
🌐 SUX Website: https://sustainableuxnetwork.com/
🎓 SUX Academy: https://sustainableuxnetwork.com/academy
📧 Get in touch: info@sustainableuxnetwork.com
