EU's New Deepfake Legislation: What Game Developers Need to Know

The introduction of the EU AI Act's deepfake transparency rules poses a unique challenge for game development studios. The legislation extends beyond the typical notion of deepfakes, encompassing a broad range of AI-generated and AI-manipulated content. While recent guidance suggests that common uses of AI in games are unlikely to fall under its scope, the distinction between fantasy and realism can be complex. The EU AI Act's transparency obligations, including the requirement to label certain AI content, are set to apply from August 2, 2026, and will cover all media types, including audio, video, and images. Notably, these regulations will apply beyond Europe, affecting studios based outside the EU that create AI deepfakes for distribution within the EU. The labelling obligation is triggered when a studio, located outside the EU, generates an AI deepfake with the expectation of it being distributed in the EU, either through incorporation in a game available in the EU or by posting the content online. Recent European Commission guidance provides clarity for game studios, indicating that common AI uses within games are unlikely to be considered deepfakes. Examples of such uses include AI voice replication for fictional characters, AI-generated fictitious environments, and AI-generated or manipulated background scenes, special effects, or technical pre- and post-processing. The deepfake obligation aims to capture content that creates a false impression of authenticity and truthfulness. However, there are circumstances where game content raises deepfake questions, such as the recreation of celebrity voices, depiction of real sports events, or inclusion of replicas of real historical figures. A high degree of photorealism increases the likelihood of such content being considered a deepfake. It is essential to remember that the obligation extends beyond the game itself, encompassing trailers, billboards, advertisements, influencer campaigns, and other social media content. The key consideration is whether the AI content could lead people to believe they are seeing or hearing something real. The AI Act defines a deepfake as AI-generated or AI-manipulated content that resembles existing persons, objects, places, entities, or events and would falsely appear to be authentic or truthful. If AI-generated content resembles someone or something that exists, can plausibly exist, or could have plausibly existed in reality, it meets the definition and could be a source of deception. Examples of content less likely to trigger deepfake labelling obligations include AI-generated voice acting for fictional characters, AI-generated fantasy worlds, AI-assisted lip-syncing or body language for fictional characters, and AI-generated fictional brands or products. In contrast, content that could be considered a deepfake includes AI-generated voice acting for real athletes, realistic AI-generated versions of real drivers in promotional materials, documentary-style game trailers depicting real-world political events through AI-generated footage, and billboards featuring AI-generated images of real musicians. When an AI-deepfake has been generated, it must be disclosed that the content has been artificially generated or manipulated, typically through the addition of a visual label. The Commission has made optional label designs available for use. The disclosure obligation has two tiers: if the deepfake is part of an artistic, creative, satirical, or fictional work, the disclosure must not interfere with the display or enjoyment of the work. If the deepfake does not form part of such a work, it must be labelled clearly and distinguishably at the time of the first interaction or exposure to the content. The European Commission's guidelines recognize that gaming imagery can fall within the first, more flexible tier for creative or fictional work. The guidance suggests a case-specific assessment rather than a one-size-fits-all approach, providing studios with flexibility. Context matters, as the same AI-generated content may be treated differently depending on whether it appears within a creative work or is used in promotional material. A sensible step towards compliance is to develop a tailored AI content policy that outlines when staff can use AI to generate or manipulate content, when such content needs to be labelled, and how to label it. This policy should be grounded in the reality of the specific studio and reflect the nature of the games they develop and the AI they use. Ideally, staff creating AI-generated content should be trained on the policy to make it a practical tool for day-to-day decision-making. The policy should lead staff through a series of practical questions to capture the nuance of the rule and avoid over- or under-labelling, including whether the content depicts a real or plausibly real person, object, place, or event, if the content is highly realistic, if audiences could believe it is authentic, and if it will be distributed in the EU. The Commission's guidance helps demystify deepfake labelling for the games industry, as many AI-powered features in games are unlikely to qualify as deepfakes. The key challenge for most studios is developing the judgment needed to determine when AI-generated content requires disclosure.