2024

System Mapping on the Environmental Impacts of AI

Systems Thinking & Changemaking

Skills

Ethnographic Research, Design Research, Synthesis and System Mapping

Tools

Figma, Adobe Illustrator and Adobe InDesign

Goal

Apply a systems-lens to holistically investigate and understand a ‘wicked problem’ using a variety of research techniques to investigate a social issue of choice and the people and communties that are directly affected. Once the research phase is complete, synthesize the information into a series of diagrams to help others make sense of the systems that exist within the challenge.

Project Overview

To build a holistic understanding of AI’s environmental impacts, I created a research dossier combining literature reviews, expert interviews, and personal fieldwork. This process helped me explore the issue from multiple perspectives — structural, academic, and human — and identify the patterns, tensions, and relationships that shape this complex system.

The deliverables included a Research Dossier and Synthesis Diagram Package.

Process

Research Dossier

The purpose of completing the research dossier was to collect as much information as possible to gain a holistic and thorough understanding of the challenge. The information collected was coded and synthesized to find key themes and form insights.

To view the full research dossier, click here: Research Dossier

Literature Review

The literature review examined existing scholarship on AI’s environmental footprint, including peer‑reviewed articles, reports, and policy documents. Across these sources, several themes emerged:

High resource consumption from data centers and model training
Limited transparency around energy use and carbon emissions
Policy gaps in regulating AI growth and environmental accountability
Social implications tied to inequitable access, digital divides, and shifting labour patterns

These insights formed the foundation for the system maps, highlighting where environmental, social, and technological factors intersect.

Subject Matter Expert Interviews

I interviewed two subject‑matter experts — Crystal Chokshi (Assistant Professor, MRU) and Nilushi Kumarasinghe (Research Associate, Sustainability in the Digital Age & Future Earth). Their perspectives helped contextualize the issue beyond academic literature. Key insights included:

Rapid AI growth is outpacing environmental policy and regulation
Energy demand will continue rising as models become larger and more complex
Environmental impacts are often invisible to everyday users
Future interventions must consider both technological innovation and societal behaviour

These conversations grounded the system maps in real‑world expertise and emerging research.

Field Work

To understand AI’s influence at a personal level, I conducted fieldwork tracking my own daily interactions with AI tools. Through journaling and data collection, I explored how convenience, automation, and habit shape my behaviour. Key patterns included:

Frequent reliance on AI for efficiency and decision‑making
Increased digital consumption tied to everyday tasks
Moments where convenience overshadowed environmental awareness
A growing tension between personal benefit and systemic impact

This reflective layer helped me connect individual behaviour to broader systemic trends.

Synthesis Diagram Package

The goal of the diagrams was to help other people make sense of the challenge and systems that exist within it through creating system maps which act as visual descriptions of how a complex system functions. The research and maps are also created based on my own interpretation and understanding of the complex and the connections within the problem.

I created the following 4 diagrams:

Problem Snapshot: Communicates the core problems that exist within the challenge.

Actor Map: Defines a network of key participants in the system and their knowledge, power dynamics, values, incentives, mental models, and relationships.

Causal Map: Defines relationships between elements that affect each other, drawing out causal relationships. This map includes leverage points which identifies where interventions might have the greatest impact.

Rich Context Map: Defines the connections between long-term trends in the system, current practices, and the emerging innovations that may effectively address some of the challenges within. This map includes some of the potential solutions.

To see the full synthesis diagram package, click here: Synthesis Diagram Package

Causal Map Process

The images below display the development of 1 of the 4 maps produced—the causal map—which showcases the relationships between different elements within the system.

First, I laid out the nodes which reflect the information I gathered in my research dossier. The nodes are organized by the following key themes: environmental impact and resource usage (green), AI policy and educating Society (blue), AI development and business practices (purple), and influence on consumer action (yellow). After drawing some connections, the first draft was printed for in-class feedback.

After the feedback session, I drew more cause and effect connections and put them into clusters. Although the map became more comprehensive it was difficult to view and derive meaningful information.

Understanding the need for better entry points and clearly showing the relationships between the parts of this systematic challenge, I reorganized the nodes into smaller cyclical clusters that connected to each other. This allowed for easier reading of the system map.

Final Causal Map

The final causal map includes a more comprehensive legend, descriptions of leverage points (possibilities of intervention), and is meant to be more scannable with different points of entry and usage of clear and simple language.

Limitations

Considering the limitations of the project being completed within the semester and the work which was shaped by my positionality, further exploration of the topic and the complex systems within would uncover more details of the issue and opportunities for change. However in consideration of these limitations, the main goals of the synthesis diagram package were the following:

Goal 1: Clearly articulate the story of the system(s) affecting your challenge

Goal 2: Provide entry points for different types of viewers (ie. the 30-second vs. the 10-minute person)

Goal 3: Ensure that your diagrams feel human and personal, not detached

Goal 4: Use principles of visual design to create cohesive, usable, and visually appealing diagrams