IT之家 10 月 1 日消息，日本计算机娱乐供应商协会（CESA）于本月上旬发布了一份有关游戏行业开发现状的初步调研摘要，完整报告将于 12 月正式对外公布。其中一组格外值得关注的数据表明：85.8% 的日本游戏开发者已经在日常工作当中积极使用生成式人工智能。
据IT之家了解，该调研的数据采集自日本游戏开发者大会 CEDEC 开展的开发者问卷调查。作为比较，去年的数据仅有 51% 的开发者正在使用该项工艺。开发者的使用情形既包含美术素材生成，也涵盖借助 ChatGPT、Copilot 这类工具处理一般性行政事务。
在已经使用生成式 AI 的受访者当中，63% 表示自己每天都会使用；22.8% 只是偶尔使用。另有 8.6% 的受访者提到，所在企业已经出台针对这项工艺的使用指引。本次调研一共回收了 1349 份开发者个人答卷。
多数受访开发者认为这项工艺可以带来益处。很多人提出 AI 缩短了制作周期、降低了开发成本；其中 837 人认为 AI 拓展了创作空间，738 人认为其降低了工艺门槛。但仅有 147 人认为，这项工艺最终改善了普通玩家的实际游玩感受。
这份调查本身也存在需要留意的局限性。问卷调研对象是开发者个人，同一家公司之内多名员工的意见会被分别统计，即便整家企业执行同一套制度；调研无法按企业维度统计总体 AI 应用情况。
而在 CESA 面向协会会员企业开展的另一项企业端调查当中，74% 的会员企业已经制定生成式 AI 使用相关制度。受访企业里面，尚未使用生成式 AI（27 家）与已经以某种形式加以运用（26 家）的数量大致相当。仅有 12 家企业将 AI 用于视觉美术素材的制作；而把 AI 运用于剧本撰写、本地化翻译、游戏测试、程序编写的企业数量就更少了。
日本游戏行业的生成式 AI 应用总体处于显著上升趋势，但更值得留意的一点是各家企业正在陆续建立严格的内部政策，以此规范 AI 落地方式。很多相关政策出台的背景，正是玩家褒贬不一的反馈以及网络上浮现的各类顾虑。
尽管在韩国、日本这类东亚国家，业界对于 AI 工艺的接受程度相对更高；不过放眼全球，类似《疯狂出租车：世界之旅》所遭遇的舆论抵制事件已经说明：游戏厂商在落地 AI 的时候必须更加审慎，不能像部分其他行业那样一味追逐生产效率提高、借 AI 缩减人力。
本文来源：https://www.ithome.com/1/009/083.htm
若有版权问题，请来信告知：13951431913#139.com 或在右方提交评论。
IT之家 reported on October 1 that Japan's Computer Entertainment Supplier's Association (CESA) released a preliminary survey summary on the current state of development in the game industry earlier this month, and the full report will be officially released in December. One particularly noteworthy set of data shows that 85.8% of Japanese game developers have already been actively using generative AI in their daily work.
According to IT之家, the data for this survey were collected from a developer questionnaire conducted by Japan's game developer conference CEDEC. For comparison, only 51% of developers were using this technology last year. Developers' use cases include both art asset generation and the use of tools such as ChatGPT and Copilot to handle general administrative tasks.
Among respondents who have already been using generative AI, 63% said they use it every day; 22.8% use it only occasionally. Another 8.6% of respondents mentioned that their companies have already issued usage guidelines for this technology. This survey collected 1349 individual developer responses in total.
The majority of surveyed developers believe this technology can bring benefits. Many pointed out that AI has shortened production cycles and reduced development costs; among them, 837 people believe AI has expanded creative possibilities, and 738 believe it has lowered the barrier to using this technology. However, only 147 people believe that this technology has ultimately improved the actual gameplay experience of ordinary players.
This survey itself also has limitations worth noting. The questionnaire targets individual developers, so the opinions of multiple employees from the same company are counted separately even if the entire company follows the same policy; the survey cannot aggregate overall AI adoption at the company level.
Meanwhile, in another company-side survey conducted by CESA targeting member companies, 74% of member companies have already formulated relevant policies for the use of generative AI. Among the surveyed companies, the number that have not yet used generative AI (27 companies) and the number that already use it in some form (26 companies) are roughly comparable. Only 12 companies use AI to produce visual art assets; the number of companies that apply AI to scriptwriting, localization translation, game testing, or programming is even smaller.
The overall use of generative AI in Japan's game industry is showing a clear upward trend, but a point even more worth noting is that companies are successively establishing strict internal policies to regulate how AI is deployed. The background to many of these policies is mixed feedback from players and various concerns emerging online.
Although acceptance of AI technology in the industry is relatively higher in East Asian countries such as South Korea and Japan, looking globally, public backlash incidents like the one faced by Crazy Taxi: World Tour already show that game publishers must be more cautious when deploying AI and must not blindly pursue production efficiency gains and use AI to cut staffing as some other industries do.