-
Episode 2 - A round of AI in games
- 2023/11/16
- 再生時間： 48 分
- ポッドキャスト
-
サマリー
あらすじ・解説
In this episode, modl.ai CEO Christoffer Holmgaard is visting DAHL's office in Copenhagen that serves as pop-up studio for our podcasts. Christoffer's music playlist turns out to be space pilot/crunch proof, and his recent played RTS game is currently on sale - I will check out playability on my Steam Deck…
We talk about applying AI in video game development. If you think the start is a bit slow, keep on listening - we will get into some really interesting corners of this topic!
The discussion is setting out from modl.ai’s own AI engine, that can create virtual players from player behavior data. This is a total different application of AI than the widespread use of generative AI, and the legal issues are also somehow different; The modl.ai engine does not use a LLM trained on public accessible data, so text- and the data mining copyright exception (TDM) is not so relevant. We are taking this further to discussing image rights and virtual players in a wider perspective than just using them for QA testing. Please enjoy.
Kindly note that the series is in English language as this is customary in the (Danish) game development community; Please bear with us when we make grammathical mistakes and tend to a low variation in expressions.
Thanks to Podcaster.dk for kind assistance with producing this episode that was recorded 27 October 2023. Also huge thanks to DAHL law firm for support, and a round of applause to Rasmus and Markus for their research.
I really appreciate feedback in any form, preferably via LinkedIn, Discord or via email to roj@dahllaw.dk.
Produced in collaboration with Podcaster.dk
Recorded by Casper Christensen
Editing and mix by Dennis G. M. Jensen/Podcaster.dk
Music by Henrik Pahlke Møller
We talk about applying AI in video game development. If you think the start is a bit slow, keep on listening - we will get into some really interesting corners of this topic!
The discussion is setting out from modl.ai’s own AI engine, that can create virtual players from player behavior data. This is a total different application of AI than the widespread use of generative AI, and the legal issues are also somehow different; The modl.ai engine does not use a LLM trained on public accessible data, so text- and the data mining copyright exception (TDM) is not so relevant. We are taking this further to discussing image rights and virtual players in a wider perspective than just using them for QA testing. Please enjoy.
Kindly note that the series is in English language as this is customary in the (Danish) game development community; Please bear with us when we make grammathical mistakes and tend to a low variation in expressions.
Thanks to Podcaster.dk for kind assistance with producing this episode that was recorded 27 October 2023. Also huge thanks to DAHL law firm for support, and a round of applause to Rasmus and Markus for their research.
I really appreciate feedback in any form, preferably via LinkedIn, Discord or via email to roj@dahllaw.dk.
Produced in collaboration with Podcaster.dk
Recorded by Casper Christensen
Editing and mix by Dennis G. M. Jensen/Podcaster.dk
Music by Henrik Pahlke Møller