Game AI: Implementation and Licensing

Demos

For an illustrative demonstration of a use case for our AI, please have a look at our reference game Marbles AI.

Also don't miss the demos and explanatory videos on our Youtube channel.

We have made significant progress since the release of these prototype demos. More demos will follow soon!

We would also be happy to give you a live demonstration of the product.

About Our AI Engine

brAIn engine is our Unity-based API for the development of innovative game AI.

We offer the integration of AI for computer games. By using our brAIn engine, we can implement orders quickly and easily and provide you with end-to-end support in the development of your games.

Unique Selling Point

Our pioneering position is based on the uniqueness of our commercial AI development interface for games. Compared to the few existing solutions with similar objectives, such as the Unity package "ML-Agents", our AI stands out for its maximum efficiency. In contrast to competitors, our AI is designed to move the training phase of the AI into the runtime of the game and to be functional in every training state. This means that the learning success of the AI can be observed by the player and thus be declared a game component. For this reason, our AI approach is also ideally suited for the implementation of gamified learning, for example for the implementation of educational games. Our AI interface also offers comparatively better integration into the development environment and the typical game development workflow and offers more extensive configuration options for different scenarios (e.g. for non-deterministic environments, real-time or turn-based scenarios, scenarios with multiple reward goals, dynamic behavioral tendencies, variable number of perceived objects, ...). In summary, our engine is better adapted to game development in different contexts and does not need to be trained in advance, which not only speeds up the development process, but also shortens the overall training time, saves resources and creates new game incentives in the process.

However, the very popular studies and demos on AI in games, such as those by DeepMind or OpenAI, which can certainly impress with their superhuman performance, have not yet been used in real-life applications in commercial computer games and are not directly comparable due to a different objective. Not only do the end devices often lack the necessary performance to calculate the AI. In particular, the learning algorithms used require exorbitantly long training times of several thousand years of playing time, so that these algorithms will probably not be able to reproduce our core competence of learning in real time in the foreseeable future. In any case, it is questionable what incentive there would be to integrate superhuman AI into regular games.

We believe that this combination of adaptability, efficiency and broad applicability makes our AI solution an international innovation.