Unviable tools?

While all this may make game AI and procedurally generated stories seem incompatible with our current education system and curricula, we would argue that there is potential for procedurally generated and game mediated stories to exist within our current system: while the system-driven narratives in question are heavily at the mercy of the AI in question, they are still confined to the previous notion of all possibilities being controlled and predefined by developers. As such, our foremost recommendation in the application of game story mediation is that the confines of procedural generation do not exceed the boundaries in such a way as to confuse learning goals and that nodular situations and points of convergence should remain consistent from student to student, and not be subject to the randomised nature of procedural generation or be trusted to computational creativity. While this may be limiting the pedagogic application of procedural generation in game mediated stories, we might still use and apply it in ways that allow the discussion between student and student, as well as student and teacher. Procedural generation might therefore be used in controlled, limited settings in order to make the differences that occur between each student’s own engagement with the game a controlled point of discussion; one such example might be the physical appearance of a character, or the layout of a room; aesthetic differences such as these are easily created via procedural generation but – when sufficiently limited and controlled – may allow these differences to be a subject of discussion in and of themselves as well as how they might act to influence the story each player experienced. One student may enter the room to find a knife, while another may find a violin, each with its own set of connotations and implications for the story. These consistent inconsistencies may not be explicitly in keeping with much of our standardised model of education but may prove to be a useful tool in developing discussion between students and teachers.

AI controlled actors that respond to the player’s decision and mediate the experienced story may allow questions of authorial intention and meaning surrounding conventionally studied texts to instead become more introspective, allowing students to question what implications their actions had upon the story and in turn what action other actors within game took as a response; such a situation blurs the line between author and reader, and may allow for the student to use the critical thinking skills and put literature into practice. Questions such as: What sort of dialogue did an AI controlled actor use as a result of student aggression, or kindness; what are the implications of the language used and why did that AI controlled actor choose that line of dialogue? Much in the same way that rules and mechanical story mediation may prove to be useful by merit of their inherently limited and predictable nature (barring bugs, glitches, and exploitation), an AI and its actors that respond to player activity and seek to guide a player’s actions within the story (either via diplomatic or non-diplomatic means) in a controlled, predictable, and limited way may produce predictable variables that engage particular problem-solving skills and engage the player in the story while conveying language and themes that are regularly taught as a part of literature focused classes. The most basic of AI controlled actors may only be able to respond with one of two options, and procedural generation may only be able to produce an environment or encounter with variables that a teacher is prepared to understand and accept. Even a rudimentary AI story mediation can open up dynamic new possibilities for schools to improve student engagement and to allow a much broader consideration of what a story is and how it might be created in an ever-changing, increasingly digital world. AI – and more broadly game mediated stories – are still in their infancy, and therefore, similarly to the complexities of social media and ‘misinformation’, it seems uncontroversial to argue that the danger of not developing new knowledge of how we might teach with and about AI-mediated gaming in schools at this juncture risks education being once again behind the literacy curve as the increasing complexity of game mediated stories and computational creativity will make the task considerably harder in the future, after the (virtual) horse has bolted.