Could AI Speed Up Car Modeling for Gran Turismo?

  • Thread starter Thread starter okami11408
  • 8 comments
  • 466 views
okami11408

okami11408

Messages
467
Singapore
Singapore
I don't have much knowledge or experience in 3D modeling, but I know AI can already help a lot with various computer-related tasks.

So I'm curious if AI could significantly speed up the process of creating car models for a game like Gran Turismo.
 
I don't think AI would be much use as it's basically copying and digitalising an already existing physical object with known parameters. There is no creative element really.

3D scanning as already massively sped up getting cars into games, just wave the laser arm around the exterior and interior then do some coding and tidy up. Back in the day they would have to sculpt the cars in CAD from blueprints or scale models, especially if they didn't have access to manufacturer information.

Now with fantasy tracks AI could possibly do something there.
 
Last edited:
Absolutely - in what ways not sure as not a 3D modeler per say. I am however familiar with 3D such as Inventor. What AI should be able to perform much faster is programming for different areas of the game.

In fact - I believe that some games released beyond 2028 should have AI capabilities built into the game. Imagine using AI in the game in livery mode to create liveries based on prompts.
AI to assist with car tunning/race settings.

I am positive that game developers that can do this will be ahead of those that do not.

I have no current plans to get PS5 or GT8. Could change, however not based on current development. GT still has not made any huge leaps as far as game play goes from prior versions in my opinion. Maybe if I had PS5 and VR with power pack might think different not sure. I do enjoy photo mode. I enjoy seeing the incredible liveries that others have time to make.
 
Last edited:
Speed might increase, but quality would suffer. I don't think it would be a good development for PD nor for the gaming industry as a whole.

The best use case for AI is to solve problems that are too complex for humans – whenever it's tasked with doing something that humans are good at, it pretty much always comes up short. Fast and relatively cheap, but of low quality compared to what a competent human being could produce.
 
eran0004
Speed might increase, but quality would suffer. I don't think it would be a good development for PD nor for the gaming industry as a whole.

The best use case for AI is to solve problems that are too complex for humans – whenever it's tasked with doing something that humans are good at, it pretty much always comes up short. Fast and relatively cheap, but of low quality compared to what a competent human being could produce.
Click to expand...
Disagree. AI can do things that humans are generally okay at doing. Much quicker and more tasks at the same time. Repetitive and reoccurring tasks are something that AI can be trained to do very well and very quick. I use AI every day at work. Especially - CoPilot with excell. As far as AI quality goes it all depends on how the paremeters and prompts are set up. You could even have things set up where AI or other AI agents are trained to check their own work. I am far from an AI expert but see the value and benefit from the value just about every day.
 
Taking scans of cars, tracks, etc and converting them into usable game code sounds like a great use of AI. I don’t see any reason for gt8 until ps6, and theres no reason for ps6 until rampocolypse is over. So yeah. Fire up that ai and give us some more tracks and cars!
 
Hopefully not because it'll reduce the need for skilled artists.
 
Keef
Hopefully not because it'll reduce the need for skilled artists.
Click to expand...
Honestly, that perspective is quite ironic, given the same people who told the physical paper pushers of 20-30yrs ago, that they were sorry..not sorry..about them losing their jobs, as things went digital, are now facing a very similar paradigm with the AI boom.

Generally speaking, of course.
 
greenlightning
Disagree. AI can do things that humans are generally okay at doing. Much quicker and more tasks at the same time. Repetitive and reoccurring tasks are something that AI can be trained to do very well and very quick. I use AI every day at work. Especially - CoPilot with excell. As far as AI quality goes it all depends on how the paremeters and prompts are set up. You could even have things set up where AI or other AI agents are trained to check their own work. I am far from an AI expert but see the value and benefit from the value just about every day.
Click to expand...
How does that disagree with what I wrote?
 
Back