Tech,  Wealth

Generative AI injects interesting soul into the game

  Recently, ChatGPT, an artificial intelligence chat tool, has become the hottest topic on social media around the world.
  When I was engaged in primary market investment, I participated in investing in a start-up AI company. At that time, the biggest problem in the AI ​​industry was that it could not find a suitable application scenario and it could not be implemented; The IQ shown in many aspects is not as good as that of three or four-year-old children, and it is nicknamed “artificial mental retardation”.
  But what is shocking this time is that the new generation of generative AI products represented by ChatGPT has lowered the application threshold of artificial intelligence a lot, and the general public can use it, so the application scenarios have increased by more than N times. . On the other hand, the level of intelligence displayed by the ChatGPT3.0 version has been greatly improved compared with the previous traditional AI products, and I believe many people have a very intuitive feeling.
  In the next 3-5 years, the generative AI represented by ChatGPT will have a great impact on many sub-sectors, especially the game industry.
Games are an excellent scenario for developing AI technology

  Since the current generative AI technology is still immature, even smart robots like ChatGPT have many imprecise and unstable situations. Therefore, generative AI will take the lead in showing its talents in the more inclusive entertainment industry.
  A few days ago, A16Z, a well-known Internet venture capital fund in the United States, issued an article saying that in terms of the types of assets involved (art, sound effects, music, dialogue, etc.), games are the most complex type of entertainment, and at the same time, games are also the most interactive. Strong, it puts a lot of emphasis on real-time experience.
  This creates a very high barrier to entry for new game developers, and also brings high costs and a long development cycle to making a truly 3A masterpiece game. These existing barriers and cost issues have created huge opportunities for the disruptive innovation of generative artificial intelligence in the game field. And “Microsoft Flight Simulator”, which is called the world’s largest simulation game, will be launched in 2022, which demonstrates the disruptive innovation brought by AI technology to the game industry. This game truly restores 2 million towns, 1.5 billion buildings and 37,000 airports around the world. It allows players to fly around the entire earth, allowing players to feel the feeling of flying an airplane in the real world. The well-known game evaluation media IGN scored a perfect score of 10.
  ”Microsoft Flight Simulator” is huge, using a total of 2000TB of map data. So how does Microsoft use 2000TB of data to model the real world? Obviously, it is impossible to rely on manual work. They chose to cooperate with a start-up company from Austria. The team has only about 50 people, but with the help of AI and cloud computing resources, they reconstructed the 3D version of the world from 2D images. 1.5 billion buildings. This is a case of a game that cannot be done without the use of AI, and more importantly, these models can be continuously improved and improved over time. In the future, game developers will be expected to use generative AI to realize some ideas that were not possible before.
  In fact, there are already many cases of using AI technology in the game industry.
  Since 2018, multiplayer online battle games that are more popular with the public and have more complex mechanisms have also become the challenge target of AI. The game AI launched by companies such as Google, OpenAI, and Tencent has played in “Glory of the King”, “StarCraft 2”, “Dota 2 ” and other complex match-up games, defeated a large number of professional teams in a row.
  Of course, the significance of game AI is not limited to defeating human players, but to make the game more fun and enhance the player’s gaming experience.
  Take the national game “Honor of Kings” as an example. As a 5V5 battle game, one teammate’s temporary hang-up will often seriously affect the game experience of others. The “AI Hosting” function launched by Tencent can simulate the similar combat power of players, and it can also cooperate with real players, which has been well received by many players.
  At present, the game industry has begun to use artificial intelligence technology in the art concept design of early game development, character synthesis and dubbing, and post-operation. With the breakthrough of the new generation of general-purpose AI algorithms represented by ChatGPT, more and more AI technologies will appear in the game scene in the future.
  The AI ​​research team repeatedly polished the AI ​​algorithm and verified the technical effect through complex game scenes, which also greatly promoted the development of AI algorithms, model structures and other technologies.
Inject a fun soul

  With the prosperity and development of the pan-entertainment ecology, digital avatars, as the core entrance of the metaverse, have taken the lead in emerging online scenes such as music platforms, games, and live broadcasts in the past two years.
  However, these digital virtual humans are more empty-looking and “human-like”, and the AIGC technology represented by the ChatGPT model is expected to connect digital humans with super AI brains and inject “interesting souls” into them , so as to have personalized expression ability.
  On February 15th, Netease announced that its mobile game “Naishuihan” under research has implemented the first game version of ChatGPT in China. This function is supported by Netease’s self-developed Fuxi AI. It will be based on the dialogue content initiated by the player to the NPC. Autonomously generate logical NPC behavior feedback.
  As a martial arts fan, I still look forward to this martial arts MMO open world game can bring some different gaming experience.
Making the metaverse world possible

  In the past two years, the fiery metaverse concept has cooled down, in the final analysis, because it is too expensive to build the content world of the metaverse through traditional methods. Through AIGC, Metaverse can meet the different content needs of a large number of users in a low-cost and high-efficiency manner.
  At present, the level design and open world construction of classic open world games such as “World of Warcraft”, “Diablo” and “Minecraft” are generated using computer programming techniques. The levels in these games are randomly generated. Every time is different, but they all follow the rules set by the level planner. Generative AI is very good at creating open worlds and level design, and it is only a matter of time before generative technology replaces procedural technology. Some startups have been seen in this field, like Promethean, MLXAR or Meta’s Builder Bot are already trying.
  There are some vast and grand world themes that would be impossible or prohibitively expensive to collaborate without generative AI tools. Just like the world’s largest simulation game “Microsoft Flight Simulator” mentioned earlier.
  In addition, it is really possible for the game to become a free and open world and allow players to get an immersive interactive experience. For example, players can freely communicate with NPCs, including text, voice, expressions, behaviors and other forms of interaction, and NPCs can independently interact with players and other NPCs in the game according to the world view of the game, their own personal settings, and interactions with players and other NPCs. With the development of the game, players will have a novel experience every time in the game, and the experience of different players is also different.
  This is what a Metaverse game should be like.
Significantly reduce game production cost and production cycle

  In an interview with the media, Lu Tangjie, head of Netease Fuxi AI Research & Technology Implementation, said that when talking with game developers who are trying to integrate generative AI into the production pipeline, what excites them most is the significant reduction in development time and cost.
  Creating great animations is one of the most time consuming, expensive and skillful parts of the game development phase. Now the generative AI model can directly transfer the expressions and actions of actors or characters in the video to the character model. Compared with the traditional solution based on motion capture, it can significantly improve the output efficiency and reduce the production cost, because it Eliminating the expense of buying expensive motion capture gear also means capturing animation from existing video.
  Another exciting aspect of these models is that they can also be used to filter existing animations.
  Another example is the AI ​​image-text generation technology, which allows planners to quickly generate and select some suitable concept maps through text, reducing communication barriers between planners and artists, and reducing the cost of intermediate rework.
  And, what’s truly transformative, these tools allow for the creation of endless variations in a highly iterative manner, in real-time, with results visible almost instantly.
  Take “Yuanshin”, the world’s most popular two-dimensional open world RPG game, as an example. It is understood that the development cost of “Yuanshin” is so high that it is expected to break the record in game history.
  MiHoYo CEO Cai Haoyu said in a speech in February 2022, “Many articles outside will say that MiHoYo spent 100 million U.S. dollars to develop “Yuan Shen” and it did cost 100 million U.S. dollars, but what they didn’t say is After the launch, we may spend 200 million U.S. dollars per year, and our development costs in the past three years will be even higher.” According to statistics,
  after the launch of “Yuan Shen” 3.0, MiHoYo’s investment in continuous development of the game will More than $500 million. At the end of 2019, the number of people in a single project of “Yuan Shen Shen” has exceeded 400. In July 2021, it was reported that the number of “Yuan Shen” development team had exceeded 1,000. Even based on the per capita annual salary of 400,000 yuan in the prospectus, the annual labor cost of the project will cost at least 400 million yuan.
  It is undeniable that Yuanshen’s large capital investment has brought about the super high quality and high operating income of the game. It is currently the most successful overseas game case in China. But such a high game investment also means that once the game falls short of expectations, it will face huge financial pressure.
  With the gradual introduction of AIGC technology into the entire game development process, 3A-level masterpieces like Yuanshen are expected to greatly reduce production costs and shorten the development cycle. At the same time, this also makes it possible for small creative game studios to develop well-made games in a short period of time with the help of AIGC technology.

error: Content is protected !!