Generative AI in Gaming: High Costs and Limited Application
- Tencent Cloud’s Liang Chen states generative AI in gaming is still experimental and costly.
- AI has been used for pattern recognition in games, but scaling generative AI remains challenging.
- Investment in AI is expected to grow annually, with hopes for more replicable applications in the future.
When it comes to the integration of generative AI, like OpenAI’s ChatGPT, into the gaming industry, the process remains in an exploratory phase. According to Liang Chen, the general manager of Tencent Cloud’s internet industry department, the use of such advanced AI in gaming is not yet widespread due to high costs and practical challenges. Chen’s insights, shared during the annual ChinaJoy gaming conference in Shanghai, highlight the cautious approach Tencent Cloud is taking towards this technology.
Historical Use of AI in Gaming
AI is not new to the gaming world. Gaming companies have long utilized AI for various functions, such as pattern recognition, to enhance gameplay and user experience. However, the leap to employing generative AI, which can understand and generate human-like content, introduces a new set of challenges. These include ensuring accuracy, maintaining historical relevance in games set in different eras, and managing the significant costs associated with training these models.
Challenges of Generative AI
Chen emphasized that while tools like ChatGPT have revolutionized content generation, they also come with the risk of producing inaccurate or illogical content. For example, training generative AI to provide historically accurate responses in games set several hundred years ago requires substantial effort and resources. Tencent uses its Hunyuan AI model and other in-house developed models to tackle such tasks, but scaling these solutions remains a significant hurdle.
AI in Game Development
Tencent has already made strides in incorporating AI into game development. The company’s AI Lab introduced the GiiNEX Game AI Engine in March, which uses generative AI to develop city scenes in games. This tool has drastically reduced the time required for 3D rendering from at least five days to approximately two hours. Despite these advancements, Chen noted that overall production time might not decrease as demand for higher quality increases and companies need to gauge user interest in the games.
Financial Performance and Future Investment
Tencent’s domestic games revenue saw a 2% decline in the first quarter, dropping to 34.5 billion yuan ($4.75 billion), while international games revenue increased by 3% to 13.6 billion yuan. Despite these mixed results, Chen remains optimistic about the future of AI in gaming. He expects Tencent’s investment in AI to grow annually, anticipating that eventually, successful applications of generative AI will emerge that can be replicated across different projects, potentially leading to a significant increase in AI investment.
While generative AI holds promise for transforming game development, its application is still in the early stages. Tencent Cloud’s cautious yet optimistic approach underscores the need for balancing innovation with practicality. As the company continues to explore the potential of AI, it aims to find scalable solutions that enhance productivity and create engaging, high-quality gaming experiences. The industry is watching closely to see how Tencent and other gaming giants navigate these technological advancements and their impact on game development and player engagement.