Generative AI in gaming: where studios are using it and what to watch

Games are among the most expensive creative products to make, with large teams and long development cycles. Generative AI promises to speed up parts of that work, and studios of every size are testing it. Bain’s survey of gaming executives found they expect generative AI to handle more than half of game development within five to ten years; most expect better games delivered faster, but fewer expect lower costs (Bain & Company).

Where are studios using generative AI?

Prototyping and concept art

Generating rough concepts, mood boards and layout ideas lets teams explore more directions before committing artists’ time. Final assets are still made or refined by artists.

Asset variations

Producing variations of textures, props and environments at scale, within a style the art team has defined, saves time on repetitive production work.

Level and world design

Procedural generation is decades old in games; generative AI adds the ability to describe a space in words and get a first draft to refine.

Code assistance

AI coding assistants help engineers write, refactor and document code and tools, and are among the most widely adopted uses.

Testing

AI-driven agents play builds to find crashes, exploits and balance problems faster than manual testing alone.

Dialogue, characters and localisation

Drafting dialogue variations, side-character lines and translations, with writers and localisation teams editing. Some studios are experimenting with non-player characters that respond dynamically to players.

Player support and community

Assistants that answer player questions from support content, and tools that help moderate chat and community spaces.

What are large publishers doing?

Major publishers including Ubisoft, Electronic Arts, Epic Games and Square Enix have described research and tools using generative AI, from dialogue drafting tools for writers to AI-driven characters and development pipelines. The common thread is AI inside existing tools and pipelines, with creative teams deciding what ships.

What are the risks?

• Rights and training data. Unclear ownership of AI outputs and of the data models were trained on.
• Consent. Using actors’ voices or likenesses requires clear agreements; this has been a major issue in performer negotiations.
• Quality and consistency. AI assets can look generic or inconsistent without strong art direction.
• Player trust. Some players react badly to undisclosed AI use. Platforms increasingly ask developers to disclose it.
• Regulation. Rules on AI transparency, such as the EU AI Act’s Article 50, cover AI-generated content and interactions with AI characters.

How should a studio adopt generative AI?

1. Start with production pain points the team agrees on, such as testing, localisation or asset variations.
2. Build into existing pipelines and tools rather than adding separate apps.
3. Set a policy on which tools and training data are acceptable, how outputs are reviewed, and how use is disclosed.
4. Keep creatives in control of style, story and final quality.
5. Measure time saved and quality, not just output volume.

Where to start

We work with media and entertainment companies on AI that pays back in production, marketing and audience growth. See our media and entertainment page, read media and entertainment marketing, or book a 30-minute session.