During this year's Taipei Game Show, in addition to showcasing many new games, the technological driving force behind the game industry—artificial intelligence, especially generative AI—has officially entered a period of explosive growth in practical applications, moving from being a "topic" last year.
During the event, Yang Shuwei, Senior Business Promotion Manager of Generative AI at AWS Taiwan, and Yang Jie, Business Manager of Netron, provided an in-depth analysis of how generative AI can permeate every detail of game development from the perspectives of cloud infrastructure and practical application solutions, from coding and art all the way to solving the "knowledge gap" and player customer service issues that cause headaches for game companies.
AWS's Yang Shuwei: Generative AI will not only draw, but also take over the "entire lifecycle" of games.
In the past, discussions about game AI often focused on whether NPCs would get stuck in walls or whether the generated artwork was natural. However, according to Yang Shuwei, Senior Business Promotion Manager for Generative AI at AWS Taiwan, the battlefield of generative AI has expanded to three major stages: game development (Build), operation (Operate), and growth (Grow).
Yang Shuwei pointed out that the current game industry is facing challenges such as a surge in content demand, difficulties in cross-platform deployment, and soaring development costs. AWS has observed over the past year that the adoption of generative AI has evolved from simple image generation to more core productivity aspects, mainly summarized into five application scenarios: coding assistance, NPC and content generation, creative brainstorming, localization translation, and quality testing (QA).
"Over the past year, the capabilities of models have become more than just an illusion; they are truly capable of solving problems." Yang Shuwei cited examples such as Amazon Q Developer, which directly assists developers in writing or optimizing code, thus overcoming the technical barriers to collaboration in distributed teams.
In terms of gaming experience, the most interesting example is Sony Interactive Entertainment's use of AWS technology to create a "player coach." Yang Shuwei shared data pointing out that "85% of players blame themselves when they fail in a game, but often it's actually a problem with the game design, or the player doesn't know how to play." By analyzing player behavior data through AI, the system can provide real-time personalized guidance (such as suggesting players adjust their crosshair position in shooting games), which is more effective at retaining players than traditional, rigid tutorials.
In addition, Amazon Games is developingMMORPG New World (America)At the same time, generative AI is also used to simulate the balance impact of adding new races (such as centaurs) to the game. This means that AI not only generates content, but also becomes a "virtual tester" for game numerical balance, significantly shortening the adjustment time before the game's official release.
Yang Jie from NetEase: Breaking down "knowledge silos" and enabling AI to become a super customer service and game analyst.
With powerful computing capabilities and models, how do game companies process massive amounts of information internally? Yang Jie, business manager of NetEase Information, points out a long-standing pain point in the game industry—knowledge silos.
Because game development involves multiple departments such as planning, art, and programming, documents are scattered across services like Wiki and Confluence, and even in the minds of senior employees. "Once someone leaves the company, there is a knowledge gap," Yang Jie pointed out incisively. The potential cost of cross-departmental communication in order to find a configuration file or fix a bug is often frighteningly high.
The NAVI enterprise-grade AI knowledge base solution, launched by NetCreative Information and built on the AWS Bedrock infrastructure, attempts to solve this problem. Yang Jie emphasized that the core of NAVI lies in security and flexibility. After all, game settings and source code are the company's top secrets. Through private deployment and access control, enterprises can more confidently "feed" data to AI.
This system has two main applications in the gaming industry:
• Internal "Game Analysts":Developers or project managers can directly ask the AI in natural language: "Where was the Boss value set in the last update?" or "What is the logic behind this code?" The AI can then act like a senior colleague who is always online and ready to respond, instantly retrieving the answers from massive amounts of documents and breaking down information barriers between departments.
• External "Super Customer Service":Unlike traditional chatbots that simply reply to generic messages, customer service representatives integrating generative AI are better able to understand the game context and provide appropriate answers. When players ask questions like "How do I pass this level?", the AI can provide accurate suggestions based on a game strategy database and can even handle multilingual requests, directly addressing the customer service manpower shortage in global operations.
Analysis: AI has evolved from a "tool" to a "collaborative partner"
In my opinion, while we were discussing whether AI-generated graphics would replace traditional art over the past year, the focus has now shifted to "how to leverage AI to streamline fragmented development processes." Whether it's AWS emphasizing the use of AI for automated game quality testing and related numerical balancing, or the enterprise knowledge base proposed by NetCreation Information, the core idea is "cost reduction and efficiency improvement."
Of particular note is that AI is making games more "understanding" of players. From Sony's AI coach to NPCs that can chat, the future gaming experience will no longer be a one-way content output, but a two-way interaction. For developers, AI is no longer just a tool for generating materials, but more like a 24/7 Copilot, assisting in handling those tedious, repetitive, or difficult-to-judge big data analysis tasks.
Of course, as AI penetrates deeper into core development processes, data privacy and security will become the next battleground. As Yang Jie emphasized, enterprise-level cybersecurity compliance will be a prerequisite for game companies to adopt AI, and this game revolution driven by generative AI is clearly only just beginning.
