Total: 1
In this position paper, we propose researching the combination of Augmented Reality (AR) and Artificial Intelligence (AI) to support conversations, inspired by the interfaces of dialogue systems commonly found in videogames. AR-capable devices are becoming more powerful and conventional in looks, as seen in head-mounted displays (HMDs) like the Snapchat Spectacles, the XREAL glasses, or the recently presented Meta Orion. This development reduces possible ergonomic, appearance, and runtime concerns, thus allowing a more straightforward integration and extended use of AR in our everyday lives, both in private and at work. At the same time, we can observe an immense surge in AI development (also at CHI). Recently notorious Large Language Models (LLMs) like OpenAI's o3-mini or DeepSeek-R1 soar over their precursors in their ability to sustain conversations, provide suggestions, and handle complex topics in (almost) real time. In combination with natural language recognition systems, which are nowadays a standard component of smartphones and similar devices (including modern AR-HMDs), it is easy to imagine a combined system that integrates into daily conversations and provides various types of assistance. Such a system would enable many opportunities for research in AR+AI, which, as stated by Hirzle et al., remains scarce. In the following, we describe how the design of a conversational AR+AI system can learn from videogame dialogue systems, and we propose use cases and research questions that can be investigated thanks to this AR+AI combination.