Pathways370
AI-Accompanied Hyperlocalized, Community-Centered Audio Walks
Sana Alimojamed, Rebeca Arens, Mei Buzzell, Maitreyi Nandhakumar, Ashish Noble
Working Prototype
https://www.figma.com/design/a8wYjmeUE4JVJhVaaCrht8/Builder-Input?node-id=0-1&t=yIYPWkSAlJA4ReGG-1
Target Audience
Any digital literate individual interested in sharing and archiving local community knowledge (ex. Place-based history).
Identified Learning Need
As physical third spaces fade and online ones take their place, the sense of community has moved into the digital realm, eroding our capacity to feel grounded and connected in the spaces we physically inhabit. Having a solid and stable connection to the space around them allows people to participate in a diversity of rich experiences, helping them form a strong local identity in their community of practice (Wenger 1999). Research has shown that a sense of community and identity fosters a deeper sense of interconnectedness (Pei et al., 2023), and the feedback and dialogue that emerges from extended interactions between community members encourages inquiry-based experiential learning (Dewey, 1938). Shared experience through interpersonal relationships and interactions, such as discussions, storytelling, and games, are key to developing critical skills such as communication and collaborative problem-solving (Panqueva, 2000). Building a community through shared immersive experiences can help cultivate a deeper sense of belonging and learning.
Rationale for AI Assistance
A physical audio walk immerses learners in the authentic spaces of the history they are exploring, fostering both situated and embodied cognition to enrich understanding and retention (Bruner, 1991). While traditional audio walks rely on pre-existing narratives, incorporating AI allows learners to actively contribute to content creation by submitting archival knowledge, personal stories, or other resources about a place. The AI synthesizes these inputs into a coherent, immersive narrative, dynamically structuring content, guiding transitions between sites, and linking stories to physical locations. This process not only personalizes the learning experience but also preserves learner autonomy, prompting reflection, exploration, and socially mediated engagement through discussion prompts or challenges (Jian, 2023).
Incorporating AI into the audio walk experience allows learners to personalize their journey by submitting archival knowledge and preferences. This personalization enables the AI to curate content that aligns with individual interests and prior knowledge, thereby optimizing cognitive load and enhancing learning outcomes (Sun & Yu, 2018). Furthermore, situating curated content within physical spaces reinforces experiential and inquiry-based learning, enabling learners to construct knowledge actively while developing a sense of belonging within the community they are exploring (Dewey, 1938). In this way, AI transforms the traditional audio walk into a personalized, interactive, and reflective learning experience that bridges archival knowledge, learner input, and embodied engagement with the environment. (Originally drafted by Beca, revised by Sana, final draft via ChatGPT)
Needfinding
Identified Learner Needs
- There is a need to strengthen people’s connection to their surroundings to support participation in diverse and meaningful experiences that shape identity within a community of practice (Wenger, 1999).
- There is a need to foster a strong sense of community and identity to enhance interconnectedness and support deeper learning (Pei et al., 2023).
- There is a need to encourage continuous feedback and dialogue among community members to promote inquiry-based experiential learning (Dewey, 1938).
- There is a need to create opportunities for shared interpersonal experiences, such as discussions, storytelling, and games, to develop essential skills like communication and collaborative problem-solving for learning (Panqueva, 2000).
Secondary Research
Literature Review
To the Castle! A comparison of two audio guides to enable public discovery of historical events (Fitzgerald, Taylor, & Craven, 2013)
A local community history group created two audio guides aimed at informing the public about a local historical event (the 1831 Reform Riot in Nottingham, England): a human-led walk that provided spoken narratives at specific points, and a technology-led walk of geolocated audio files that provided similar information through a GPS-enabled smartphone. They sought to evaluate the differences between the two and if they were an effective learning experience. They found that participants on the technology-led walk liked the independence of the tour and the technological affordances, particularly the clearer listening experience and ability to play back the audio, and overall displayed similar levels of learning as the human-led walk. However, participants preferred shorter, less drier content, missed the opportunity to ask questions on the walk, and wanted different versions of the audio recordings to accommodate the listener (i.e. native vs non-native English speakers, children vs adults, levels of familiarity with the history). The researchers reached the conclusion that technology-led tours are a viable, scalable approach for community history groups to foster informal learning opportunities about local history, and we believe that AI could be leveraged to address user issues and support the aforementioned approach.
The effect of audio tours on learning and social interaction: An evaluation at Carlsbad Caverns National Park (Novey & Hall, 2005)
Researchers evaluate the effect of an audio tour, using casual narration and ambient sound, on visitors at Carlsbad Caverns National Park. They compared audio tour users and nonusers on a 12-item knowledge quiz while evaluating other behaviors such as reading informational signs and interacting with one another, and found that audio tour users’ scores were roughly similar to the nonusers’ scores on fact-recall multiple choice, but that audio tour users were more likely to answer open-ended questions on the park’s main themes correctly. Audio tour users and nonusers’ social/interpretive behaviors (reading signs and interacting with others) did not differ substantially, though at audio-only stops nonusers were more likely to be interacting with other visitors. Overall, the audio tour was received positively, increased the length of the visit, and did not strongly impede social interaction while fostering deeper learning about the overall message and themes of the park.
How the Use of Audio Walks as Creative Public Engagement Expands Access to Site-Based Heritage to a Diverse and Globalised Audience. (Reagan, 2024)
Reagan argues that the digital format of audio walks offers greater audience reach and flexibility compared to in-person tours, despite the drawback of lacking real-time dialogue with visitors. The chapter further analyzes the demographic data of the audio walks, revealing that a large percentage of listeners engage with the content remotely rather than taking the walks in situ, suggesting the content functions more as an “audio experience.” Ultimately, Reagan encourages heritage sites to prioritize digital engagement offerings, recognizing that visitor interaction may differ significantly from initial expectations.
Memoryscape: How Audio Walks Can Deepen Our Sense of Place by Integrating Art, Oral History and Cultural Geography. Geography Compass. Butler, Toby. (2007).
The article explores the history, practice, and theoretical implications of creating sound walks or “memoryscapes”. It argues that audio walks serve as a powerful platform for sharing knowledge and amplifying diverse narratives within a community. They can be a vehicle for oral histories, allowing long-time residents to share their memories and perspectives with newer generations. This form of storytelling can be particularly impactful in preserving and disseminating the cultural heritage of a neighborhood.
Existing Commercial Solutions
Google Audio Walk Creator
https://artsandculture.google.com/experiment/talking-tours/8AGlfzgsYmBeIA?hl=en
Google Audio Walk Creator allows users to choose specific streetviews at cultural landmarks, and generates audio insights.
- Location-Specific Analysis: Google’s Gemini multimodal language models analyzes Street View images, GPS, and historical context.
- Script Generation: The AI audio model creates real-time audio tours, and allows users to generate new ones.
- Query-Based Interactivity: To learn more about a space, users can choose from 3 pre-generated questions or ask their own.
Echoes
https://echoes.xyz/
Echoes enables creators to build free or custom AR audio experiences through tailored apps, including locative audio walks, guided tours, or immersive sound experiences.
- Spatial Mapping: Echoes offers geolocated audio playback, and uses live heatmaps to trigger a dynamic experience based on location
- Surround Sound: Supports stereo, binaural, 3D, and ambisonic audio
- Branching Narratives: Adapts story progression to the listener’s movement, direction, and spatial location, making it a a mechanism for narrative choice
VoiceMap New York City
https://voicemap.me/tour/new-york-city
VoiceMap NYC is a platform that allows users to access self-guided walking tours delivered through local storytelling, routed by real-time location on a map.
- Geo-Synced Narration: GPS-based automatic audio narration is synced to listeners’ position through a mobile app.
- Collaborative Storytelling: Audio tours are contributed by a diverse community of journalists, artists, and locals, using soundscapes and narration to create immersion.
References:
- Bruner, J. S. (1991). The narrative construction of reality. Critical Inquiry, 18(1), 1–21. https://doi.org/10.1086/448619
- Butler, T. (2007). Memoryscape: How audio walks can deepen our sense of place by integrating art, oral history and cultural geography. Geography Compass, 1(3), 360–372. https://doi.org/10.1111/j.1749-8198.2007.00017.x
- Dewey, J. (1938). Experience and education. Macmillan.
- FitzGerald, E., Taylor, C. & Craven, M. To the Castle! A comparison of two audio guides to enable public discovery of historical events. Pers Ubiquit Comput 17, 749–760 (2013). https://doi.org/10.1007/s00779-012-0624-0
- Jian, M. J. K. O. (2023). Personalized learning through AI. Advances in Engineering Innovation, 5(1), 16–19. https://doi.org/10.54254/2977-3903/5/2023039
- Novey, L., & Hall, T. (2005). The effect of audio tours on learning and social interaction: An evaluation at Carlsbad Caverns National Park. Science Education, 91(2), 920–935. https://doi.org/10.1002/sce.20184
- Panqueva, G. (2000). Play, puzzles and creativity: Learning engines for the knowledge society (Technical Document ACE-01-00, Version 2, p. 452).
- Pei, L., Poortman, C., Schildkamp, K., & Benes, N. (2023). Teachers’ and students’ perceptions of a sense of community in blended education. Education and Information Technologies, 29(2), 2117–2155. https://doi.org/10.1007/s10639-023-11853-y
- Reagan, R. . (2024). Unlocking Heritage Stories: How the Use of Audio Walks as Creative Public Engagement Expands Access to Site-Based Heritage to a Diverse and Globalised Audience. In C. Wergin & S. Affeldt (Hrsg.), Digitising Heritage: Transoceanic Connections between Australia and Europe (S. 73-89). Heidelberg: Heidelberg University Publishing. https://doi.org/10.17885/heiup.1305.c18420
- Sun, J. C. Y., & Yu, S. J. (2018). Personalized wearable guides or audio guides: An evaluation of personalized museum guides for improving learning achievement and cognitive load. International Journal of Human–Computer Interaction, 35(4–5), 404–414. https://doi.org/10.1080/10447318.2018.1543078
- Wenger, E. (1999). Communities of practice and social learning systems. Organization, 7(2), 225–246. https://doi.org/10.1177/135050840072002
Team Contributions
- All team members ideated together, then did preliminary research and ideated some more independelty;
- Beca wrote the original sections;
- Ashish, Maitreyi, Mei, and Sana provided feedback on edits;
- Sana updated the description based on a group discussion. Rationale for AI Assistance was edited by chatgpt.