Storya’s Manifesto on AI and Creativity (Part 1)
We set down our current stance on the collision of Big Tech’s AI and human creativity, and propose guidelines for creators’ use of AI.
The first question to answer is: why us?
We have spent the last four years researching, experimenting and exploring the intersection of Artificial Intelligence and creativity. We approach this topic as fiction and non-fiction writers, painters, digital artists, tech founders, and more.
We have a unique perspective, as both artists and tech practitioners, of where this powerful technology will take humanity. We are not trying to be the much-despised “moderates” in this conversation, but we believe Artificial Intelligence is here to stay, and our role is to figure out the correct way to use it to unlock the next frontiers of human imagination.
But that is not the only motivation driving us. We want to build a world less controlled by established narratives and power structures. We want a world where storytelling in all its forms flows freely across borders and cultures, connecting us all. We believe technology has a role to play in achieving that, alongside the need for a far greater focus on how we bring diversity, equity and inclusion into play in the creative industries at large.
There are stubborn bottlenecks stifling creators around the world, and especially in the emerging markets of Asia and Africa: some due to antiquated market infrastructures (e.g. traditional publishing), some due to uneven availability of technology, others because of cultural, economic and political oppression and exploitation, which takes many forms around the world.
So, why call it a Manifesto?
Because we believe the conversation around AI has reached such a fever pitch that it is more important than ever to establish practical but reasoned guidelines for AI in creativity. While creators should and will continue to push boundaries, we believe it is key for all, starting with our own creative practices, to come up with a set of guiding principles for how to think about, and use, AI.
Many of our positions are critical of the corporate and tech interests behind AI, while some of them will be hard for the Luddite-inspired artists to accept, too.
In short, we are trying to upset as many people as we can!
Jokes aside, we do not have all the answers. But we believe we have some good questions to ask and some work in progress answers based on our on-the-ground experience. This Manifesto is just another stone on the path to our collective, nebulous future, but having the conversation take place is critical.
The Three Key Questions:
Given the rapidly growing range of AI technologies capabilities in the realm of creative work, we want to structure this document (which we will publish in parts) around the three fundamental questions of the AI debate.
- Is generative AI capable of autonomous creativity?
- Is genAI a threat to artists’ livelihoods?
- Is genAI a threat to human creativity?
Based on the answers, we will propose our Principles. But, before we dive in, we must set the scene with some context.
Creators vs Big Tech
During our creative journeys, AI companies unleashed a new generation of technology unto the world, with two tools in particular achieving incredible popularity in a matter of months: ChatGPT by OpenAI and the image generator MidJourney. The result of releasing these new technologies has been a deeply divided world.
We have the tech companies insisting that this is all for the best. Not just the best for their wallets, which have grown considerably, but for humanity at large. We also have the same tech companies insisting that they are the key to our salvation, should this tech get out of hand and unleash Armageddon. Quite a confusing start!
On the other hand, we have artists who hate AI and everything it represents.
- Fiction writers wonder if they will be soon obsolete, while also wondering why AI is so terrible at creative writing. Several, mostly commercially successful, writers are suing tech firms like OpenAI for taking their works without permission to train their large language models.
- Visual artists hate what image generators represent, a form of cheating that robs art of its very essence. Some go as far as calling it machine “vomit”. And, like writers, they resent the theft of their works to train models that can now produce art in their style.
- Voice artists and actors are worried about their jobs going extinct, as audiovisual avatars already show remarkable ability for emotional range.
- Filmmakers, while least affected, for now, see the fast advancing of text-to-video technology as a signal that Hollywood may be next in the line of fire.
And this is but a partial list. Collectively, many organizations, from unions to publishers’ associations, are pointing the finger at Big Tech and AI as THE biggest threat to their very existence.
The next piece of context is around the tech itself.
What “art” can this technology “create”, right now?
With more or less complex text prompts, AI tools already allow one to generate a very broad range of creative output. These include:
- Any form of creative text, from short stories to poems, from song lyrics to full length novels or even interactive stories.
- Photorealistic or artistic images in any current or past style conceivable, from anime to cave paintings, on any subject, from the most abstract, to the most mundane, to the outright toxic (deepfakes). Text-to-video is one way to do it, but AI can equally interpret rough sketches and render them into fully realized “art” in real time.
- Audio at song length that successfully mimics a broad range of musical styles from any era. Music videos are on the list, too.
- Short videos that can be micro stories with characters, dialogue and settings.
- Animated avatars (any avatar, from oneself to the Mona Lisa to a fictional character) to recite any text from point 1. For the avatars, all that is required is a single image of the avatar’s desired appearance.
- One can code entire villages of autonomous AI agents that can engage in free-flowing conversations with each other, without additional human prompting. We can literally watch machines create or simulate virtual communities. This has taken some fascinating forms, such as a social media site that is entirely populated by AI agents, no humans allowed (except as spectators).
- In video games, AI can take any non-playing character (i.e. a computer-driven character previously limited to pre-loaded actions and dialogue) and transform them into evolving, natural-sounding characters in the story world to allow players a level of immersion unseen until now.
There are more use cases that are not as clear cut, namely the use of AI as a copilot across existing and new categories of creativity, and we will address those as well. On those, there seems to be some high-level agreement that usage fenced around one’s own content should not be discouraged.
“If you’re using AI to create derivative works of your own work, that is completely acceptable,” said Authors Guild CEO Mary Rasenberger.
That is, unfortunately, where the niceties end, and controversy begins.
More in the next section of the Manifesto.
Stay tuned.