Rising Costs of AI Tokens Concern Game Developers

As generative artificial intelligence (AI) becomes more prevalent in the gaming industry, the expectation that AI will replace entry-level developers is proving to be misguided. While replacing developers with AI is challenging, the rising costs of token usage are prompting a reevaluation of cost-effectiveness.
On June 24, the Chief Technology Officer (CTO) of a gaming company stated, "While the introduction of AI has indeed improved work efficiency, it is not yet at a level where it can replace human workers. It is difficult to compare costs directly with personnel."
In the game development field, the role of skilled developers remains crucial. Although generative AI can assist with coding, debugging, and documentation, complex problem-solving during game engine design and large-scale service operations heavily relies on developers' experience and judgment.
The industry currently views AI more as a productivity-enhancing tool rather than a means of workforce replacement. It is seen as a way to help developers reduce repetitive tasks and focus on more critical responsibilities.
Recently, game companies have expanded their support for AI tools, including ChatGPT and Gemini, with the addition of Claude, which is favored by developers. A major domestic gaming company has implemented Claude and Gemini Enterprise across the organization this year. Claude Enterprise is reported to cost a minimum of $250 per person per month.
At the current exchange rate, this amounts to approximately 380,000 won. The company supports all employees, regardless of their roles, in utilizing AI, allowing them to apply for over 18 different AI services based on their job functions. Considering the company has around 9,500 employees, the associated costs are significant.
Other game developers are also providing enterprise-wide support for Claude. They operate a system where development teams can apply for the specific AI tools they need.
Currently, the productivity gains are offsetting the increased costs, but as AI usage rises, questions about efficiency are emerging. An industry insider noted, "While token costs are still cheaper than labor costs, if the current trend continues, we will face a situation where token costs exceed labor costs."
Another Chief AI Officer (CAIO) in the IT sector pointed out that major AI service providers are shifting from flat-rate to usage-based pricing models. He explained, "In the past, many contracts were subscription-based with fixed costs for a certain period, but recent enterprise agreements increasingly involve additional charges based on usage. The more AI is used, the higher the costs."
He added, "Companies with previous contracts could predict costs over a set period, but recent contracts link increased usage directly to cost increases. Large corporations are increasingly grappling with finding a balance between expanding AI usage and controlling expenses."

* This article has been translated by AI.