cross-posted from: https://lemmy.world/post/40224994

I just found out that Embark used genAI to replace recalling it’s voice actors. And I am devastated. It was my GOTY and now I’m boycotting it. To the point that I decided to be vocal about it.

I don’t want any AI (not talking about machine learning for the way the Arcs move, I’m talking about the voice actors being replaced) in the work of art that I think are the video games I play.

To game developers: Stop using AI in your video games!

Je viens de découvrir que Embark a utilisé de l’IA générative pour remplacer ses acteurs. Et franchement je suis tellement déçu. C’était clairement mon GOTY, et maintenant je me retrouve à boycotter tout ça, au point de carrément en faire des posts pour toucher le plus de monde possible.

Je ne veux aucune IA (et attention, je ne parle pas du machine learning utilisé pour les déplacements des Arcs) dans les oeuvres d’art que je consomme.

Aux développeurs: Arrêtez d’utiliser de l’IA dans les jeux vidéos !

  • Holytimes@sh.itjust.worksEnglish
    0·
    17 hours ago

    I have played enough games that have lasted enough time to know that it’s not uncommon to be unable to get a voice actor to come back for many reasons up to and including death or retirement. Which results in worse experiences for the player because now there’s just a sudden voice change which is unpleasant. Or your developers are now limited because of scheduling another issues which causes problems.

    For them to be able to have a consistent voice throughout without having to rely on that in the voice, actor only has to do a one-time gig get paid in the potentially earn a small annual stipend from the project depending on their contract.

    Genai for voices is really really good in that case. You can sell the use of your voice to train a model for a small lump sum with a yearly licensing fee or a large up front sum.

    Easier in a more reliable paycheck for the voice actor. As long as the contract is fair and compensation is done correctly and the licensing is on a strictly project by project basis. So no using that trained AI from multiple projects.

    Like this is probably the single best example of using genai that we can really hope for. It could be a slight bit better. Sure, but this is already leaps and bounds ahead of anything else.

    Bespoke case by case trained AI with contractual agreements for payment in licensing.

    And the best part of all of this since it’s contractually agreed, it means no one is forced to do s***