Much like the technology itself, the applications of artificial intelligence in games have evolved. Decades ago, video game AI tools used to be specifically designed for each game and were largely reserved for hard-scripted behaviours, such as Pac-Man’s four coloured ghosts. Non-player characters’ (NPCs) behaviour was created utilising ruled-based systems and finite state machines. Today AI tools have progressed towards learning methods technologies, (like generative AI), which are used to make NPCs' behaviours more complex and lifelike, resulting in a better game experience.
Generative AI technology is a machine learning method that uses large data sets to teach AI models patterns and structures. This wide range of data includes text, images and computer code taken from a variety of sources like the internet through web scraping, access to database and open source repositories.
Creating AI environments to train models that are robust enough to generate truly smart and fully independent agents (one that takes actions autonomously and acts intelligently, learning in order to achieve its goals) is difficult and time intensive. Open source AI tools have emerged offering environments for training intelligent agents using deep reinforcement and imitation learning like the Unity Machine Learning Agents Toolkit (ML-Agents). These types of tools have quickly become popular with games developers.
Using open source AI tools offers undoubted benefits, but also carries risks that need to be properly addressed to ensure the long-term future of the game relying on them.
Open source content can be used without royalties and/or explicit consent. This is not the case with most other publicly available content which makes open source content an attractive alternative or often the only option for those in need of content for AI tools. The content is generally made available subject to the user’s compliance with the terms and limitations set out in the applicable open source licence. Open source licences can take different forms and be subject to different levels of restrictions. It can be time consuming to work through the terms of each relevant licence which means developers will find using open source a challenge during a development process with tight deadlines.
In most cases, the obligations related to the open source content will be triggered by the distribution of the game incorporating the open source elements. These obligations are likely to include requirements to provide specific information such as retaining the copyright notice, attributing the author or identifying modifications, retaining certain clauses of the open source licence, indicating the origin of the repositories, and supplying the work that incorporates the open source content under the same conditions as the open source licence.
In addition, open source content is usually made available without any guarantee or possible recourse against the creator, publisher or distributor of the software. Any risk of infringement of third party rights would be fully borne by the user.
How to reduce risk
To mitigate the risks of using open source AI tools in games, consider:
These relatively simple steps can save you a great deal of trouble when you start distributing your game.
Erik Steiner looks at the opportunities in the games industry offered by AI, and at how to manage the associated legal risks.
1 von 6 Insights
Desideria-Alexia Pohl looks at how to help ensure ethical use of AI in games.
2 von 6 Insights
Martijn Loth highlights the top ten considerations to help games businesses mitigate risks associated with using generative AI when developing video games.
3 von 6 Insights
Our international team looks at copyright issues around the use of generative AI in video games.
4 von 6 Insights
Laura Craig and Miles Harmsworth look at the use of personal data in AI tools used by the video games sector, and at the evolving regulatory framework.
5 von 6 Insights