(This is a series of short ‘previews’ of papers to be presented at the upcoming Experimental AI for Games workshop at AIIDE 2014. Tune in live on Twitch on October 8th to catch the presentations of these papers, or find the PDFs online at http://www.exag.org)
‘The Ideas Person’ has a bad reputation in the games industry – someone who offers up game concepts but doesn’t want to pull their weight. But everyone needs ideas from time to time, and when we’re stuck for inspiration, maybe it wouldn’t hurt to have a source of ideas on-hand? In their paper ‘Towards the Automatic Generation of Fictional Ideas for Games‘, Maria Teresa Llano Rodriguez, Simon Colton, Rose Hepworth, Michael Cook and Christian Guckelsberger describe their ‘What-If Machine’ (WHIM) project and how it might be applied to invent ideas for games. Here’s a preview.
The paper brings in ideas from WHIM, a computational creativity research project to build a system that can automatically generate ideas, assess how good ideas are, and then invent simple examples to show off the ideas in different domains. This paper discusses what an idea generator might mean for the games industry – both players and designers. The authors begin by discussing fictionality and the many ways in which it influences game design, but the later parts of the paper introduce the What-If Machine in its current state, and talk about two ways it might apply directly to games.
One way in which WHIM is currently able to generate ideas is by stringing together and modifying knowledge from ConceptNet, a huge online database of real-world concepts. ConceptNet holds information like ‘camels are animals’ and ‘animals can hear sound’, stored in a logical format that software can easily parse and manipulate. By connecting together facts like these and searching for particular kinds of relationship, WHIM can propose interesting questions that could serve as the basis for a song, a story, or a videogame. ‘What if there was a little cat who was afraid of milk?’ is just one example from the paper.
How can this help make better games? The authors suggest two ways in the paper. The first is that a system that can generate ideas is a great thing to have to spark new ideas, inspire extensions to existing ideas, or to help people fill in the gaps when they’re struck with creative blocks. WHIM’s existing templates can suggest things including possible characters, superhero abilities, and unusual situations, all of which could be applied during a game design. They even include an illustrative example at the end of the paper, where two 15 year-old students used WHIM’s output as the basis for a sketched game design; a really nice example of how the system could have real uses for designers of all ages and experience.
The more intriguing application for WHIM is to actually include it in the game as a mechanic or system that can work with or against the player to improvise new ideas and scenarios. The authors liken it to improvisational storytelling – the player declares that they have a sword to go adventuring with, but WHIM can use its knowledge to rewrite the story so that the sword becomes blunt, leaving the player to think of a new way to solve the problem. Embedding WHIM into a game could be an opening for a strange and unique multiplayer experience, where Player Two is a computer.