Creativity AI #6: Your weekly productivity and creativity insights
Midjourney office hours (5th Feb 2025), Can you copyright AI images, Ideogram text tool, Perplexity new models, AI photo festival
Creativity AI is a new publication that brings together interesting articles and recent developments in the AI world related to creativity and productivity.
Midjourney Office Hours (2025-02-05)
Midjourney V7 is still in its training phase. There is no confirmed release date as of yet.
The current V7 model is based on an older dataset. A new, larger dataset will be included in an upcoming update. The enhanced dataset will better capture exotic details (such as mythological or historical elements).
Existing features such as personalization, moodboards, and Sref codes will continue to be supported.
The future V8 version may add complexity for further enhancements.
The developers are still debating how to incorporate video and 3D into the overall workflow with limited resources.
Can you copyright AI images?
The US Copyright Office has released a report that addresses the legal and policy issues surrounding AI and copyright. You can download a copy here.
Overall, the document emphasizes the importance of human authorship and creative expression as key elements of copyright protection.
This is a 52-page document. Like many other busy people on the planet, I don't have time to read through a complex legal document in detail. I only have time to browse through it casually and look at the pictures/figures.
So I ask Perplexity for help in summarizing and tabulating it WITHOUT HALLUCINATION. How do you do that? Request that it quote the exact phrase to support the key facts it generated.
Here is the result:
The simplified guides below are provided solely for informational purposes. Consult a legal attorney for current and professional advice.
What Can Be Copyrighted
Original human-drawn artwork that is input into an AI system and remains clearly perceptible in the output
Creative modifications and edits made by humans to AI-generated images
Original human-authored text combined with AI images in a creative compilation (like a comic book)
The creative selection, coordination and arrangement of AI-generated elements by a human author
Human-authored expressive elements that are retained and visible in the final AI-modified output
Human-created derivative works that creatively transform AI outputs
The overall human-authored composition of a larger work that incorporates some AI elements
Human artistic choices in modifying specific regions/sections of AI art
Human-authored elements in a mixed work containing both AI and human-created content
What Cannot Be Copyrighted
Content generated entirely by AI with no human creative input
The underlying AI-generated elements of a modified work
Simple text prompts used to generate AI art
Random variations generated by AI from the same prompt
AI-optimized or AI-expanded versions of human prompts
The selection of a single AI output from multiple generated options
Ideas or concepts described in prompts to AI systems
The technical process or method of generating AI art
Why does this matter?
Even if you don't intend to "register" or claim the copyright of your AI art, understanding what can be copyrighted will keep you from accidentally using copyrighted materials in your project (and potentially getting into trouble).
It appears that post-production work on AI-generated images (with human involvement) could be considered "value-added" because it may assist the AI artist in claiming copyright. I'm not sure if this is a good thing in terms of sharing free AI art with the community. However, it might make people wary of using AI images from the internet as is because no one knows if those images have been extensively edited or altered by a human.
Ideogram
Ideogram has released its new Text Tool, which allows users to overlay text on images on Canvas. Users can also use an external image to add and edit text.
It functions similarly to any text editing tool found in other standard graphic software, but there is no blending mode.
Why does this matter?
Users can add a large amount of text directly to the images. There is no need to switch to another graphic editing software for text editing. Nothing fancy, but we can’t add text directly like this in Midjourney.
This feature makes it easier to create personalized greeting cards, book covers, and more.
Perplexity
Perplexity has updated its list of AI models available. The new OpenAI O3 Mini model has also been added.
Why does this matter?
If you don't like the answer generated by an AI model, simply rewrite it using another model. The results can vary greatly depending on the model you've chosen. Essentially, you pay a single subscription fee and gain access to all models.
You can also tabulate the results using various AI models offered by Perplexity to see the common key points covered by the models and their citation sources.
AI Photo Festival, anyone?
The Almagine exhibition in Brussels (January 24th to June 15th, 2025) features 18 artists who use AI to revisit and reimagine historical events, characters, or situations.
Quite interesting, hein?
If anyone tells you that AI art is not art, send them - the arty-farty folks the link to the exhibition website so that they can buy tickets and see it for themselves.
Inspiration
I come across a beautiful illustration by Ernst Haeckel.
And I used it as the only image in a Moodboard to create similar artwork. Guess what: it works even with just one image!
Intriguing medieval illustration created using a Moodboard code (schkhhl).
Recent articles
Free Geeky Animals' Sref code: 3896264700
Cover prompt: A person walking away from a labyrinth of intertwined hearts, leaving behind confusion and embracing clarity, cinematic lighting, high detail, symbolic storytelling --ar 16:9 --sref 2652678841 --v 6.1
Invitation to publish on Creativity AI
Please contact me if you want to publish on Creativity AI.
This publication will be distributed through Substack and Medium platforms, reaching thousands of readers.
Email: animalsgeeky@gmail.com