Assessing the Impact of AI on Development: Separating Fact from Fiction

The debate over the use of generative AI in game development has become a contentious issue in recent years. Engaging with this topic can be challenging, as it often sparks intense reactions from those who view AI as either a revolutionary tool or a morally and ethically flawed concept. Lewis Packwood has noted that AI is already being utilized in various aspects of development, although this is often kept under wraps to avoid backlash from consumers. However, the question of whether AI can truly accelerate development and to what extent remains uncertain. The morality of AI, particularly with regards to the vast amounts of stolen work used in its training, is a complex issue that raises personal and legislative concerns. Nevertheless, for businesses, the practical question of whether AI delivers on its promises is of paramount importance. The lack of comparative case studies and best practices due to the secrecy surrounding AI use in many studios creates an information vacuum that can be exploited by unscrupulous vendors. The problem AI claims to address is indeed existential for many studios, as they struggle with increasingly unmanageable development cycles. The promise of AI to significantly boost productivity across various aspects of studio operations is naturally appealing to studio heads seeking to regain control over their timelines. However, while AI can be beneficial in narrow, supervised applications, its capability to autonomously deliver substantial productivity gains, as touted by AI companies, is far from the reality. On the programming side, AI can be useful as a tool for skilled programmers, acting as an advanced autocomplete feature that saves time and repetition. This supervised use of AI is how most developers are leveraging the technology today. It can offer moderate productivity gains, especially in the prototyping phase, but it is far from the paradigm of autonomous AI completing tasks independently. The actual bottleneck for studios remains the need for skilled, experienced coders, who are costly and difficult to find. Similarly, in artwork and other fields, while AI might aid in prototyping by speeding up the process and enhancing demo quality, the assets produced are often not consistent or of high enough quality for inclusion in final products. The consensus is that AI-generated art, while sometimes impressive, lacks the reliability and quality required for shipped games. The benefits of AI in these areas are debatable, particularly considering the additional workload it can create for artists to fix AI-generated flaws. Generative AI does not solve labor shortages; studios still need to hire and pay skilled artists to correct and refine AI-produced assets. The core issue that studios aim to solve is the scarcity of skilled labor, which has been a persistent problem in the industry. The appeal of AI lies in its promise to finally solve this skills gap, but this promise is often based on dubious claims and a lack of understanding of how AI scales to more complex problems. The failure to recognize the limitations of AI technology may prove disastrous for early adopters, leading to project failures and potential studio closures. Consumer backlash is a lesser risk compared to the financial and operational damage that can result from over-reliance on unproven AI solutions. Eventually, AI will find its place in development, but its impact will be more limited than its proponents suggest, and some studios will have to learn this lesson the hard way.