What's "Cat Royale"?
Cat Royale is an interactive installation where three cats (Ghostbuster, Pumpkin, and Clover) spent 12 days in a custom environment while a robot arm played games with them. An AI system learned which activities each cat preferred and adapted its play strategies accordingly. The artists created what they called 'a utopia for cats' to explore care relationships between animals and autonomous systems.
You can find more Cat Royale videos and highlights from each day here.
Who's it for?
Cat Royale is aimed at public audiences interested in AI and care, as well as researchers studying human-animal-technology interactions. The project is also relevant for policymakers and technologists working on the regulation of autonomous systems.
Approach

The project uses Research through Design with an interdisciplinary approach. Artists, computer scientists, and animal welfare experts collaborated to design and build the AI system. Over 12 days, the system learned about cat preferences through observation, while welfare officers continuously monitored to ensure ethical treatment. The approach demonstrates how rigorous welfare protocols can be integrated into experimental AI research.
Contribution

Through the design and deployment of Cat Royale, the project demonstrates that measuring and optimising 'happiness' algorithmically is far more complex than technology companies suggest. The work shows how to engage animal welfare experts and the public in designing autonomous systems from the outset, thereby creating a model for responsible AI development. In addition to practical contributions, such as welfare protocols and public engagement methods, the project provides theoretical insights into the nature of care, trust, and the limitations of algorithmic decision-making.
Why is it in the Observatory?
This project offers a concrete example of how Design Research can make AI ethics accessible and engaging for public audiences. By using the familiar context of caring for pets, it demonstrates how to facilitate meaningful conversations about algorithmic decision-making, trust, and responsibility in ways that technical papers or policy documents cannot achieve.