Microsoft debuted in the Muse, the generative AI model, which can be trained to lead visual video games and predict the input data of the controller. The company describes this technology as the future of the idea of the game, not something that can be used instead of human creativity during the development of the game.
Microsoft has published his discovery of Muse in a science journal Natureand published two Blissful ia video About this. It is possible that you can still have questions about what Muse is doing. Unlike the most popular AI generative models that are infamous Steal work -protected work with copyrightMuse was built on the player’s data from The edge of bleedingA delicate shooter at the Ninja Theory arena, which appeared in 2020 and was abandoned only 10 months later. The information collected from the game allowed Microsoft to train Muse on over 1 billion photos and controller’s activities, in the amount of seven years of continuous gameplay.
The resulting AI model is able to produce faximile (though ugly) The edge of bleeding The game, which, according to Microsoft, is proof of Muse, has “a detailed understanding of the world of 3D games” and can “create a coherent and diverse gameplay.” Chairman and General Director Microsoft SATYA NADELLA shared the film on x By irritating the possible addition of the Muse to Copilot, Chatbot from GPT-4 with GPT-4 in Windows 11. Phil Spencer, general director of Microsoft Gaming, even painted the future in which Muse can facilitate the behavior of video games.
“You can imagine a world in which data and video from the game could learn old games and really make them portable to any platform on which these models could work,” said Spencer. “I think it’s really exciting. We talked about protecting games as an activity for us and these models and their ability to completely learn, like a game without having to start the original engine on original equipment that open a lot of possibilities. “
House Matthews, head of the Ninja Theory Studio, also quickly separated the earnest and obvious fears of Muse, which has already pushed the professional workforce in the video industry from the picture.
“Such technology is not about using artificial intelligence to generate content, but it really is about creating work flows and approaches that allow our team here 100 creative experts here to do more, go further, and faster toe in a tangible form. I think that an interesting aspect that is exciting is how we can use such technology to make a process of games faster and easier for our talented team so that they can really focus on what I think is really unique in games, So this human creativity – said Matthews.
Talking with New scientistHowever, AI researcher and game designer Mike Cook said that Muse is “long, far from the idea that AI systems can design games themselves” due to inherent training restrictions in one game.
“If you build a tool that actually tests your game, running the game code itself, you don’t have to worry about perseverance or consistency because it works in reality,” Cook explained. “So they solve the problems that generative artificial intelligence introduced.”
Even if human creativity is not threatened in Muse, as Microsoft says, it is complex to balance hearing about this developing technology, knowing that many people are still grabbing many mass dismissals of Microsoft. The presentation of technology such as Muse as experimental or somehow completely favorable to creativity is unthreatening and simple when its output looks like nonsense, which at this moment does unequivocally (not to mention, she is trained in one game). We only hope that Microsoft will keep this word when generative AI models improve to such an extent that complex results of programmers.
