How generative AI is changing creative workflows

It's the end of 2022, and this year has been to some extent revolutionary for AI. Well, at least in the media. There has been plenty of buzz around generative AI models. In simplicity, these models are primarily here to create "new" content. I put new in quotes because many people don't realise two things.

First, these methods have been around for some time. It's just that someone has been brave enough to invest an enormous amount of computation and labour to achieve the results we are seeing on social media.

Second, the content these models create isn't entirely new, but a reflection of our own creativity. Machine learning models use our generated data to learn how to create new content, but they aren't able to go beyond that.

The definition of creativity is the use of imagination or original ideas to create something. That doesn't mean these models aren't useful. History has shown that such technologies can be disruptive to an industry. Take professional translation. It's not that long ago we saw similar excitement about neural translation models among practitioners. While those models promised increased efficiency and higher profits, the reality was that many top-level professional translators saw their margins decrease as they were viewed more as proofreaders than as creative professionals. It is important for us as AI professionals to be mindful of these limitations and to understand the potential impacts of these technologies on different industries.

Generative AI will also be part of game development. Like it or not. The adoption has accelerated in the last two years. Just look at the SIGGRAPH conference and how many AI-related papers are being submitted. There are potential applications in the creation of game content, such as levels, characters or dialogue. Another exciting topic is the creation of personalised game experiences. A generative model could create unique scenarios or challenges based on a player's skills, preferences and playstyle.

I believe the game experience is what we should focus on when we bring these powerful models to the game industry. To me, this is about giving creative people tools to go faster and make more iterations. Kapnetix is creating models for motion capture because digitising movement accurately makes a great game experience. We first need to automate the non-creative work so that the creative work can have focus and stand out. That belief is reflected in our mission:

Our mission is to help VFX professionals focus on creativity and create the experiences we fall in love with.