How AI will affect the next generation consoles?

We left 2018 behind with great developments in the technology world. 2018 is a year in which the future of the game world is gradually determined, where processor architectures are down to 7nm, the year we saw great improvements in processors’ power consumption, GPUs have real-time Ray-Tracing processing and great development in machine learning. In the first days of the new year, I started my research to explain how the next generation of consoles that we think we will see in 2020 can be affected by these developments. This article will focus on how the development of machine learning and artificial intelligence will affect the new generation consoles.

PlayStation 5 and Xbox Two

First of all, I would like to tell you the short history of artificial intelligence. Artificial Intelligence, which we often encounter in everyday life as AI, is the ability of a computer to perform different activities in a way similar to intelligent living things. First of all, Alan Turing, who is the father of computer technology, once thought, “Can computers think?”, took the first steps in the development of the artificial intelligence and he was the one who encrypted the German communication codes which is called Enigma during World War II. This first step was really important because it was showing that computers could help people. Turing started a new era by breaking the Enigma and he is one of the biggest geniuses of history that allowed computers to come to this day.

The reason why artificial intelligence has gained so much place in the agenda in recent years is the developments in Artificial Neural Networks. In connection with the number of transistors that have reached incredible numbers in proportion to the downsizing of the architecture of the GPUs and processors that is the brain of our computers, we now have computers that can simulate our brains in a sense. Here we need to open parenthesis and explain the ability of people’s thinking and learning. The vast majority of our intelligence is obtained by trial and error. For example, if we touch something really hot, our hands burn. When our brain encounters such a situation, it processes the information it perceives to establish a connection between the neurons and learns that we should not touch the glass when it is hot. If we want to take a cup which is hot, we think of a different method. The method we try here can be using two layers of cups, and when our brain encounters the same condition, it sends a signal to neurons which this information is processed and it uses two glasses and prevents our hands from burning. Artificial neural networks teach artificial neurons by trial and error method just like this. When they encounter a problem, they learn the solution of the problem by using this information.

This is the basis of the technology that allows phones with artificial intelligence to work faster or take more beautiful pictures. If you take your phone when you wake up and immediately look at your e-mails, the phone will keep track of these activities during the days and at the end, it gives you this information in a much faster way because it knows what is your routine. Such examples can be shown as evidence that Artificial Neural Networks are now can be seen in today’s life.

How GPU’s helping AI development?

Nowadays, processor and GPU manufacturers offer extra processing units for processors and GPUs which are produced specifically for Machine Learning. These processing units for machine learning are able to process and get a result of multiple data simultaneously in a parallel way. Although standard CPUs can be replaced by multi-core processors, they can never reach the same performance as the GPUs with thousands of cores. In particular, Nvidia’s CUDA kernels’ machine learning capabilities are unquestionable. The rapid use of GPUs in artificial intelligence training has also begun to change the image of video cards as game-oriented hardware.

Of course, with the development of artificial intelligence in parallel with the graphics cards, will begin to see the impact of this technology in games. Especially with the new generation of consoles, we will start to see the effects of artificial intelligence more clearly. Now let’s examine what will be the effects of this technology on the gaming side.

What will AI brings to next generation consoles?

We have addressed the subject of machine learning in the article. We are already seeing the first effects of machine learning in games. You must have heard the news about the improved quality of photos and videos with the help of artificial intelligence. Using machine learning, we are able to increase the resolution of low-resolution photos in the most appropriate way. This technology has been started to be applied to old games using different software. For example, with the help of ESRGAN (Enhanced Super-Resolution Generative Adversarial Networks), the textures of old games can be remade without any human touch involved. Morrowind is one of the games which use this technology. Although we can’t say that the graphics are new generation, it can compete with the remaster games. Image-quality enhancement methods with artificial intelligence support can form the basis of the next generation. How? Like this:

As you know, one of the biggest handicaps of PlayStation 4 was that it didn’t support previous generation games. The reason for this that the architectures of PS4 and PS3 were very different from each other, as Sony said. However, we will probably not have the same problem between PlayStation 4 and PlayStation 5. PlayStation 5 is most likely using AMD Ryzen CPU and GPU with Vega infrastructure, so it uses X86 basis, it is getting easier for old generation games to work on new generation of consoles. Nevertheless, players will not only play their old games in the next generation of consoles but also they will expect higher resolution and graphics quality.

The ESRGAN method I have just mentioned can be integrated into the next generation of consoles so that the consoles can improve their resolution and enhance the texture quality with 4K resolution. Because developers are not included in any way, we’ll be able to play with both better texture quality and higher resolution without waiting for the Remaster version. Personally, I think such a feature can be great for the next generation of consoles.

How will AI improve our gaming experience?

Another potential effect of developments in artificial intelligence is, of course, that the characters with artificial intelligence will become much more realistic. I think the new generation will also stand out with their living worlds as well as its graphics. With more advanced artificial intelligence, we’ll be seeing much more advanced versions of NPCs of Red Dead Redemption 2. This is very important for RPG games and will allow games to have a much more realistic atmosphere.

Artificial intelligence can have great innovations in terms of user experience other than these two. New generation consoles are likely to come with Siri-like assistants. In this way, even if your console is off, with the help of a voice command, you will be able to start a download and perform many different activities. This feature has been tried in Microsoft’s Kinect, but it has not been successful. With the updates and innovations on artificial intelligence, there is a chance that it will work more successfully.

Of course, apart from the ones we have mentioned, we will be able see the innovations made with creative ideas of the developers. Personally, I think that artificial intelligence will be one of the basic features of the new generation of consoles. Although these developments seem to be quite new, the performance increase rate of artificial intelligence is so high, so that I think these will shape the future of the consoles.