Conversations With A Tree
Project Details
Overview
Tools
Python, Raspberry Pi,
Inworld AI & Chat GPT
Timeline
Apr - May 2025
"Conversations With A Tree" is an interactive installation that connects communities with the environment through a chatbot embodying a native tree's persona. This immersive educational experience shares the tree's cultural, historical, and ecological significance, including its ecosystem role, indigenous traditional uses, and conservation efforts. The chatbot personalizes responses based on users' interests, making the tree's story accessible to those who communicate verbally. By merging technology with nature, the project fosters deeper understanding of biodiversity and environmental stewardship.
Introduction
The project began with creating a series of storyboards, system diagrams, and dioramas of various ideas before a concept was selected as the primary idea. Although many storyboards depicted different ways of a person having a conversation with a tree, the final outcome of having a person directly talk to a chatbot attached to the tree became the final result. Various conversation methods between humans and trees were explored through multiple storyboards, but ultimately the design settled on implementing a chatbot interface affixed to the tree that would enable direct communication.
Storyboards and dioramas
Method
After defining the project scope, the required hardware and software were obtained. The chatbot was built using InWorld.AI, which lets you create custom characters with specific knowledge and personalities. The chatbot encourages users to ask about events it has witnessed, its environmental role, traditional oak uses, and other questions.
Proposed oak tree for installation
After creating the profile in InWorld.AI, a Raspberry Pi was equipped with a microphone and speaker, and the character's API was connected to the Pi. Libraries for voice and listening were also installed on the device. An environment was created for the API and libraries so that whenever the Raspberry Pi was turned on, the program would work without the need to install the libraries again.
System diagram for conversational tree & flow
Documentation
To document the project's work, a mock test was conducted on the morning of the final presentation, where the device was attached to a tree with a nearby outlet on campus. Signage was also created to show users what questions they could ask the chatbot and what type of tree it was personifying on campus.
Two students conversed with the chatbot during the testing phase. During the first round of testing, the first user was excited to talk to the chatbot and asked about how it was used as a medicine by indigenous peoples of the local area. This user also appeared delighted by the chatbot's answers and touched the bark of the tree and looked at the branches while the chatbot verbally gave out information.
Conclusion
Conversations With A Tree offers a way for community members to learn about trees they encounter daily but might not know much about. Despite some technical challenges, the installation showed that conversational AI can make plant knowledge more accessible.
Next Steps & Further Development
-
Refine chatbot prompts to be more specifically geared towards the tree's identity as a native plant, preventing users from asking overly general questions
-
Improve physical device design by creating a wooden case that mimics the specific bark of the tree, better integrating the device with the natural environment
-
Address user experience issues by updating signage and chatbot prompts to explicitly tell users to ask about the specific tree itself, helping manage expectations around response timing
-
Optimize technical performance to reduce the 5-10 second latency that confused users during testing, though this may require additional development time and resources