Unity’s new AI-driven tools and Marketplace news is another indicator of controversial technological advances in the video game industry. Developers of all sizes using Unity will soon be able to use text prompts to create game content and build games that can integrate AI models.
According to Marc Whitten, president of Unity’s Create division, this technology and tools from other providers will be “disruptive” to the video game industry because such technology is clearly so powerful. In a public statement and interviews with game developers, Whitten outlined his vision for Unity’s AI tools to enable “creators” of all kinds to create their 3D content at a faster, higher quality pace. described as a natural evolution of the company’s long-standing mission to help
We wanted to know more about how Unity is tackling themes like data security, rapidly generated ‘clumps’ of low-quality AI content, and backlash against games like . life best and sky Utilize AI technology for development.
Whitten answered our questions and shared insights on Unity’s AI efforts.
Some of Unity’s AI technologies have been in development for years
The question many developers are asking generative AI tool makers is, are they operating on their long-standing practice of creating game development tools, or are they more recent investors in tools like ChatGPT and Midjourney? Is it riding on the enthusiasm of
Whitten explained that Unity has been experimenting with tools like machine learning and generative AI for years. The ‘animate’ tool within Unity Muse allows animators to interpolate the frames of their animation, and he’s one of the tools Unity has been working on for years. The AI model inference tool Unity Sentis is built on Unity’s Project Barracuda, which was first released to developers in 2021.
But it was advances in natural language input, embedded paint, and the sub-adaptation process that inspired Unity to think about a “fundamental platform to connect” its various machine learning tools.
Questions about using large data models quickly turn to the sources of the data these models extract. Currently, tools such as ChatGPT and Midjourney face criticism because they rely on datasets that collect content from the internet.
Whitten declined to summarize the exact sources of Unity’s datasets, but provided some background on what developers can expect in terms of data ownership. He said Unity Muse Chat is built on “other big language models” outside of Unity that the company has licensed, and “heavily tuned” around Unity’s own information. In doing so, developers trying to troubleshoot Unity issues will be able to interact with a chatbot interface whose responses are weighted by their knowledge of Unity’s documentation.
Regarding concerns about content scraping of visual assets, Whitten said Unity is “very focused” on building an “ethical and transparent” model around these processes. Our goal is that when a developer creates an asset using Unity Muse, that asset is included in the full game license so that the developer or your studio can take full ownership of the content. is to
“We make sure we build them on models that we have the right to and have a clear understanding of where the data comes from to make sure they work,” he said. Stated.
AI models can now produce increasingly reliable output when trained on ever-evolving datasets. This has prompted some tool makers such as Adobe to draw new image data from platforms such as Adobe Stock. Practices like this have made some creators worry about their content being used to train AI models without their consent.
Whitten said Unity improves its models by adding more data, but said the company’s focus is not “taking other people’s art and training on top of it.” “We are empowering creators to take control, understand what is going on, be transparent, and give people the right choices,” he said, powering Unity Muse. I wasn’t able to confirm whether the model I’m working on would be automatically trained on content created by Muse, but I suspect that automatically designing the system to do so would “break” the promise of data transparency. repeated.
What does Unity think about the backlash against AI?
As mentioned above, the introduction of generative AI into video games has met with considerable backlash from fellow developers and players alike. Witten declined to comment on the following examples. life best and skyHe said most of the developers he’s interacted with so far are excited about the technology’s potential, and that speeding up workflows and “democratizing” the game development process is at the heart of Unity’s service.
“I think the line between AI and non-AI tools is very blurry,” he said, citing vegetation modeling middleware SpeedTree as a comparative example. “Nobody using SpeedTree is designing the leaf-by-leaf tree that appears in-game. It’s about filling it with vegetation that looks good.”
“That’s exactly what we do with our generation tools.”
Whitten said that in working with developers, the company has mostly seen use cases for AI tools where developers rapidly “ping-ponge” between generated content and manual adjustments.
What Whitten was less willing to comment on is that low-quality AI-generated “chat” is starting to fill various markets. Developers working with AI tools may have two different markets competing with each other, as AI games rapidly created on Steam can have similar performance impacts on mobile game stores. , creators make a living by creating middleware and assets for the Unity Asset store. You may be attacked by similar spam.
Whitten said there is a “deeper question” on the topic of AI spam that game developers are having a harder time finding players, even if their games are of very high quality. Stated. He said this is an issue that Unity is trying to help developers with through Unity’s Grow division, and that the company will always take the time to “listen to the people putting content on the Asset Store” with developers. He said it’s been spending a lot of time and definitely improving the experience.
Unity’s comments and bragging about the power of generative AI are unlikely to allay the concerns of developers who are loathing the technology. The company’s language about Unity Muse is so similar to the boasts I’ve heard from other execs that it’s easy to imagine the backlash.
But when it comes to platform AI technology, Unity Sentis could be the ace of this platform. Whitten is a development interested in using access to neural networks (minus the higher cloud computing costs for him) to perform tasks such as automating the day/night cycle and creating broader responses to NPCs. He said he had already spoken to him.
Developers seem enthusiastic about creating games with large-scale AI models that enable new and unique experiences, and a reliable and cost-effective way to integrate these models is a concern for some developers. to the Unity ecosystem. This will certainly benefit Unity’s business as a whole. model.
