3.11 Applying AI to your environmental project

This module transitions from conceptual understanding and case study analysis to practical application. It is designed as a guided, worksheet-style exercise to empower learners to conceptualise their own small-scale, AI-assisted environmental project. The goal is not to produce a finished project within the scope of this course, but to develop a well-considered, actionable plan that builds the confidence to take the next steps in one’s own community. The process is broken down into five key planning stages, each referencing the frameworks and tools discussed in previous modules.

1. Defining your question: The power of specificity

The most successful community projects begin not with a grand ambition, but with a specific, answerable question. Instead of “How can we stop pollution?”, a more effective starting point is a focused inquiry rooted in your local context.

• Actionable prompt for learners: What is one specific environmental question or challenge in your community that you are passionate about addressing? Frame it as a clear, manageable question.
• Examples:
  • “Where are the most common illegal dumping sites along our local stretch of the river?”
  • “Which native pollinator-friendly plants are disappearing from our local parks?”
  • “What are the main concerns our community’s elders have about changes they have seen in the local climate over their lifetimes?”

2. Identifying your data: What stories can you tell?

With a clear question, the next step is to consider what kinds of data could help answer it. This data can come from a wide variety of sources, both traditional and digital. This stage involves combining different types of information, such as oral histories from residents, community reports, and scientific measurements.

• Actionable prompt for learners: Brainstorm all the potential sources of data that could relate to your question. Think broadly about both qualitative and quantitative information.
• Examples (linked to the questions above):
  • For illegal dumping: Geotagged photos taken by community members during walks, official municipal sanitation reports, oral accounts from long-term residents about historical dumping spots.
  • For pollinator plants: Photos of plants in local parks, historical planting records from the municipality, interviews with local gardeners or botanists about changes they have observed.
  • For elder concerns: Recorded oral history interviews, analysis of local newspaper archives, historical weather data from a nearby station.

3. Choosing your tools: Matching the technology to the task

Not every problem requires the most complex AI. This stage involves selecting the appropriate AI approach based on your question and your available data, referencing the tools and techniques discussed in Modules 3 and 4.

• Actionable prompt for learners: Review the tables “AI-Powered Citizen Science Tools” (Module 3) and “AI Techniques for Environmental Data Analysis” (Module 4). Based on your data types, which one or two AI tools or techniques seem most suitable for your project?
• Examples:
  • “If my primary data is photos of illegal dumping, I could explore using a citizen science platform where users can upload and map these photos. I might later use computer vision to try and classify the types of waste in the images.”
  • “If I have interview transcripts from community members, I will use an NLP tool for thematic analysis to identify the most frequent concerns.”
  • “If I want to track pollinator plants, I will start by using a tool like iNaturalist to identify and map observations.”

4. Applying your ethical framework: Building on a foundation of trust

Before a single piece of data is collected, an ethical framework must be in place. This is a critical step for any project involving community knowledge or personal data. While this course does not explicitly contain a dedicated ethics module, the principles of data sovereignty, informed consent, and benefit sharing are crucial for building trust and ensuring your project is respectful and effective.

• Actionable prompt for learners: Outline how you will address key ethical considerations in your project plan.
• Examples:
  • Data sovereignty: “My project will establish a community data agreement, stating that all photos and stories collected belong to the community group, not to me as an individual.”
  • Informed consent: “I will draft a simple, clear consent form for anyone I interview, explaining exactly how their stories might be used, including for AI analysis, and I will have it reviewed by community members first.”
  • Benefit sharing: “The final output of the project – a map of illegal dumping hotspots – will be shared freely with the local environmental committee and the public to advocate for more cleanup resources.”

5. Planning for the “human-in-the-loop”: Where expertise matters most

AI is a tool, not a replacement for human expertise, judgement, and context. This final planning stage involves identifying the critical points where human oversight is non-negotiable.

• Actionable prompt for learners: Identify at least three points in your project workflow where the “human-in-the-loop” is essential.
• Examples:
  • “A human expert must verify all AI-generated plant identifications before they are added to our final map.”
  • “The thematic analysis from the AI will only be a starting point. Our project team must conduct the final interpretation of the community’s stories to ensure we understand the cultural context.”
  • “Before we publish any AI-generated visualisations, they will be reviewed by the community members who provided the data to ensure they are accurate and respectful representations.”

By completing this structured planning process, learners will have transformed a general interest into a concrete, ethical, and technologically informed project concept, ready for further development.