OpenAI and Studio Ghibli style: Theft or homage?
This week's episode of Modem Futura explores how OpenAI's new image generator is stirring up old questions about art, authorship, innovation, and a whole lot more.
In the latest episode of Modem Futura, Sean Leahy and I dig into the ethical and legal landscape around AI-generated images, spurred on by the recent Studio Ghibli style controversy ignited by OpenAI’s newest image generation model.
As we note in the podcast, people have been grappling with the challenges as well as the opportunities of AI-generated images some time (including whether “art” is an appropriate description here). Yet as AI image generators get more sophisticated — and especially, as is the case with OpenAI’s latest offering, they are capable of mimicking styles that are synonymous with human artisanship which represents hundreds of hours of skilled work, the issues raised are becoming more pointed and more complex.
Underneath the usual banter, this was a nuanced conversation around an issue that is far from black and white.
And of course, I couldn’t resist abandoning my usual use of Midjourney version 3 to use ChatGPT 4o to generate the image for this post! The image above is based on a transcript of the podcast and a photo of Sean and myself in the recording studio, and using a style of ChatGPT’s choosing.
I have a sneaky idea that ChatGPT didn’t appreciate the irony of the caption, given the image rendering 😊
Listen below on Apple Podcasts or on Spotify or YouTube. And if you want to jump to specific entry points, check out ChatGPT o1-Pro’s generated summary with approximate time stamps.1
Entry points from ChatGPT o1-Pro:
AI image generation and its rapid evolution
Andrew and Sean discuss significant advancements in generative AI, particularly focusing on image generation platforms/models like DALL-E and Midjourney. They highlight how these AI models have evolved rapidly from creating unrealistic images to nearly photorealistic ones.
Intellectual property and ethical dilemmas in AI art
The conversation explores ethical and intellectual property issues raised by generative AI. Sean and Andrew discuss how AI-generated art challenges existing legal frameworks and ethics around ownership, copyright, and the use of human-created works as training data.
The Studio Ghibli controversy
The podcast addresses the controversy surrounding OpenAI's new image generator explicitly capable of replicating Studio Ghibli's distinctive animation style. The discussion explores the moral implications of recreating iconic styles without permission, potential legal issues, and the broader implications for art and creativity.
Vibecasting! (and human vs. machine creativity and artistic value)
Sean and Andrew explore the differences between human and AI-generated creativity. They focus on the intrinsic value of human artistic processes and discuss whether machine-generated art diminishes or complements human creativity and effort.
Fan art and public ownership of creative works
Approximate timestamp: ~01:02:56
The podcast explores fan art, AI-generated images, and the concept of public ownership once a creative work is widely embraced. The hosts question how creators' intentions align or conflict with how the public engages with their works, particularly in the age of AI.
As always you can watch the episode on YouTube as well as listening to us (and this image was generated using Studio Ghibli style):
If you find the podcast useful or think others will enjoy listening, please do share it with colleagues and leave us a rating or review.1
Thanks!
Andrew
These are the themes that OpenAI o1-Pro thought were important. Lightly edited.
Thanks for this episode. This is an important, complex topic, one that I'm still trying to think through. Coincidentally, the issue is relevant to a doctoral seminar I'm teaching on information systems ethics, security, and privacy. Also, I LOVE the artwork, especially the RODE on the boom arms! That's awesome (I'm a fellow podcaster)!