do you see any future where game development is completely reliant on AI? or do you think it'll progress naturally as a tool or "forbidden"/looked down on in studios? internally, not externally
I don't think we'll ever be "completely reliant" on AI. Right now AI is a tool that we're still evaluating - if we can iron out the legal and cost issues with AI and it significantly increases work efficiency, we will probably incorporate it into our workflows. This wouldn't really be any different from using an off-the-shelf engine like Unreal or licensing middleware tools like Havok for physics simulation. It's something we may choose to use when the tool can provide for a specific need and the cost to benefit ratio is deemed worthwhile.
There are some elements of AI that some developers are using already internally, such as using a LLM to index and act as a search engine for internal documentation. Smaller games don't need to do this, but long-running games with lots of systems and content in place often have thousands of pages of documentation that needs some way to sort through. LLMs are one of the easier and more efficient ways of organizing all of those data. Some designers and producers are using AI to generate art for internal pitches and presentations. Most of this is kept to internal systems, we can't risk intellectual property being pulled into the greater AI databases as training material because of the potential legal issues there.
Ultimately, it comes down to companies evaluating a value proposition. If AI costs too much, doesn't do enough, or is too big a liability, we won't use it. If AI can do the work we need it for at a reasonable cost with minimal liability, we'll start seeing more developers use it.
[Join us on Discord] and/or [Support us on Patreon]
Got a burning question you want answered?
Short questions: Ask a Game Dev on Twitter
Short questions: Ask a Game Dev on BlueSky
Long questions: Ask a Game Dev on Tumblr
Frequent Questions: The FAQ
