top of page
  • Michael Jonsson

AI tools: why we shouldn't be afraid of bad content!

Creativity in a new technological era.



The recent explosion of AI tools for digital creatives left many people skeptical about the quality of content that could soon follow. It seemed obvious that there would, at some point be a deluge of easily produced, low quality content on the Metaverse. Of course there would! Every time the barriers to creating something are lowered, anyone with fingers can essentially make the thing. Artists are afraid not only of losing their jobs but of being lost in a deluge of mediocracy.


...Until they realize what it actually signifies: the democratization of creativity, specifically which was, until now behind a fairly high difficulty wall. And for those with inspired ideas, creative flare and talent, it now becomes that much easier to bring a vision to fruition with far less friction.

The metaverse needs content! Its future health and vitality depend on it, the easier it is to create, the more interesting and diverse of a place it will be. Which means anyone can find content that resonates with them. Sometimes people just want an odd idiosyncratic experience rather than a polished corporate space, which is exactly what art actually is: a view into the mind and unique perspective of another individual.

Creator-centric platforms and tools are democratizing the ability to create immersive content and experiences, meaning anyone can create and share their own content, regardless of their technical skill or budget. This low barrier to entry will lead to a flood of new voices and perspectives, making the Metaverse a more creatively diverse and interesting place.

It can also serve as a stepping stone for creators to learn and improve on their skills as its hard to create the type of content that inhabits the metaverse, such as optimized 3D assets like avatars, digital fashion, entire scenes or environments or interactive objects and experiences. Skills such as 3D modeling, texturing etc are hard to learn and take time. It can be frustrating for someone with a vision but lacking the technical background to go about creating it. There are many creators out there who have wonderful ideas but are unable to share them, which means we all miss out.



Light Tree: AltspaceVR installation art.



I personally can attest to this process, when I first started world-building, I was uploading Tilt Brush scenes onto AltspaceVR (and still do). The program I was using wasn't an AI tool per se, but it was easy and intuitive, I could create an entire virtual world in a fairly short space of time, that was also as close to the vision I had in my mind as it was possible to be: it's painting in 3D after all. Point is, I was able to let my imagination flow without too many constraints and was therefore able to deliver easily accessible, magical and immersive experiences for anyone to enjoy. Had this simple, easy to use tool not been available to me, that content wouldn't exist, people wouldn't have been able to experience it. It also provided me with the confidence and the stepping stone to continue my world-building journey, adding skills and tools along the way, from additional 3D modeling and texturing skills, then learning how to build/ upload for various different platforms.



Poetry and Magic: new AltspaceVR world created for the virtual burning man (BRCvr).

World link: https://account.altvr.com/worlds/1505600991771230884/spaces/2173310804093305838



If we keep these tools difficult to use and the assets difficult to create, we will end up with a paucity of interesting content and experiences in the Metaverse both good and bad, as well as a lack of creative diversity. This is not good for adoption or decentralization! Do we want the overlords that be, to dictate the experiences we can have or create? We also want the ability to vote with our content, meaning that if a platform is is pursuing unethical or nefarious actions, we won't just take our eyeballs elsewhere, we can take our content also! Content after all, is the lifeblood of any platform: people will go where there are interesting things to see and experience. Most of these assets are easily portable across platforms and if creators own their own assets, they have agency over where and how they are displayed.


In a recent interview with ZDNET, Neal Stephenson - who originally coined the term 'Metaverse' in his book 'Snow Crash' and is now overseeing the creation of a layer1 blockchain: Lamina1, the substrate for the open metaverse, said: "I think some of the empowerment of creators that was lost in Web2 we may be able to recover in Web3 if we structure these things in a way that's more aware of the requirements of individual creators,"


Will there be a lot of 'bad' content? Yes, but it's a small price to pay for creative freedom. It also doesn't necessarily mean the metaverse will be filled with nothing but the mediocre, there will always be well crafted, polished experiences by professional artists and world builders who continue to push the boundaries of immersive experiences, the same as there are classically trained oil painters as well as hobbyists who happily coexist. But my intuition tells me there will be more good than bad: putting easy to use tools in the hands of talented creators will only increase the output of inspired content.


Poetry and Magic: new AltspaceVR world for BRCvr.



Although this all does come with certain caveats, especially when it comes to text to image generation AI tools: some of these tools are trained on data sets containing the work of living artists, there are some obvious ethical problems with this and we couldn't write this article without attempting to address them. It's not the ease of use which is considered unethical, it's the ease at which another artist's style or work could be recreated - in this case. So, as a rule of thumb it's best to not to evoke the name or style of a living artist and where possible, even consider training your own models with photographs you've taken or iterate on your own original art, you might generate some unique and unexpected results!



This image and the one below are originally created pieces that have been iterated on by Michael Jonsson using AI tools.




And finally a message to the artists and creators who are worried about their futures: keep creating! The world always needs original and unique voices and visions! Use some of these tools to your advantage to enhance your skills and amplify your creativity. Here are some 3D/ world building AI tools to check out if you're an aspiring 3D artist:


1. move.ai - motion capture from mobile video.


2. kaedim3d.com - Image to 3D model creation - try generating a 3D model from a drawing or photographing an unusual object then create a 3D model from it.


3. inworld.ai - character generation.


4. anything.world - interactive 3D experiences and animations.


5. Nvidia Omniverse - Nvidia's full content creation hub that optimizes workflow between various 3D creation software.


6. Metaverse platforms with AI tools built in for creators such as Monaverse who have an AI image generator for texturing built into their unity toolset and Simulacra who have AI driven 3D creation tools.



The last word: If artists give up, we are giving up on art itself, on originality. We might in the end, only be left with models trained on their previous output in an infinitely recursive loop if that plays out to its ultimate conclusion. It's also possible that in future, we will need artists specifically to generate original output for this reason, in which case, they should be paid by the companies and platforms who develop these models and gather the training data to keep the input data fresh, a ubi for artists if you will - now there's an idea!!

Simon Willison and Andy Baio actually came up with an open source search engine that combs through image training data sets, collected by LAION and used by Stable Diffusion. Which means artists now have the ability to see if their work has been used in the training data. Simon Willison wrote a detailed post about about the search engine. "His Datasette project is open-source, extremely flexible, and worth checking out. If you’re interested in playing with this data yourself, you can use the scripts in his GitHub repo to download and import it into a SQLite database."

Read the full article here: https://waxy.org/2022/08/exploring-12-million-of-the-images-used-to-train-stable-diffusions-image-generator/





If you'd like to build with us or come on our podcast, get in touch here: https://www.reachxr.io/book-online



Hazel Griffiths



0 views0 comments

Recent Posts

See All
bottom of page