How to Use Generative AI in Environmental Research and Consulting:  A syllabus

About this syllabus

This course was designed for environmental science, environmental studies, and sustainability majors at the university where I teach.  I created it because I saw that while some students were using AI, many were just using it as a fancy Google or a homework-answering machine, sometimes in ways which would undercut their learning.  Meanwhile, my colleagues in the private and public sector professions were using generative AI for all kinds of purposes, from ideation to coding, and had begun thinking of their work in terms of managing agents.  I was concerned about whether we were preparing students to join this type of workforce.

After teaching the course, I have come to think of this less as course about "AI", and more of a course about how to think.  I believe that anyone with basic critical thinking and data analysis skills can build interesting things with AI.  Key point: AI education is not solely about learning particular "tools".  To get good output from a LLM, you need to be asking the right questions. 

Many colleagues have heard about this course and said, "I wish I could take a course in how to use generative AI!"  So I wanted to put it online — with the caveat that this was purpose-designed for undergraduates who are interested in the environment.  This course is slightly revised based on experience from Spring 2026.

Course description

Are you interested in how generative AI can help you do research and work on projects?  Are you interested in the environment and sustainability?  This class is designed for total beginners to intermediate users of generative AI.  We’ll start with the basics of where this technology came from, how it is evolving, and knowing when to use this tool and when not to use it.  Then we will explore the different uses cases of generative AI - including summarizing, writing, designing graphs, creative applications, quantitative applications and writing code.  We will do this with an eye to projects you might do if you were working in research or consulting, through guided and self-selected examples that relate to environmental challenges in energy, resilience, conservation, and more. We will wrap up the class by zooming out to take a big-picture look at the environmental, social, and ethical implications of different futures for this technology. Students can expect to leave the course equipped to responsibly use generative AI in work that relates to environmental studies and sciences or in other fields.

What are we going to learn?

By the end of this course, you will be able to…
  1.  Understand the fundamentals of generative AI and be familiar with different generative model types.
  2. Identify and act on opportunities to responsibly use generative AI in your own research and consulting work, as well as identify times when it is not the right approach for the task.
  3. Critically evaluate AI-generated output to identify biases, factual and logical fallacies, missing context, and limitations.
  4. Evaluate the strengths, current limitations, and future possibilities of Large Language Models from a technical standpoint.
  5. Understand the range of environmental, ethical, and social issues that have been identified with regards to generative AI, and develop your own ideas on these, enabling you to meaningfully weigh in on debates around AI policy and governance as well as its application in environmental sciences and studies.
  6. Build a project that showcases your skills in generative AI use to future employers or clients.

Course flow

The first part, “Foundations”, is about basics.  It aims to make sure we’re talking with a common language, and that we understand the history and context of the tools we’re going to work with.
 
The second part, “Applications”, is designed around all the different things you might want to use generative AI for in your work and life.  Many of us are already using tools like ChatGPT or Claude for things like answering questions, developing outlines for papers, or summarizing readings.  We will take the use of these tools further by learning how to use generative AI to work with data, make predictions, write code, and more, working through examples from the environmental social sciences, ecological science, and environmental humanities.
 
The third part, “Futures”, brings us back to a wider look at what this all means — for our work, society, and the environment.  We’ll discuss how different schools of thought think about the future AI, and the options for policy and governance to guide how it develops.
 

Readings and recommended media

This course is partly scaffolded around Ethan Mollick's 2024 book Co-Intelligence: Living and Working with AI.
The landscape is always changing, which means no reading list will ever be totally current.  Recommended reading and listening includes:

Course schedule


, PART 1.  FOUNDATIONS

Week 1:  What is generative AI, and where did it come from?

Objectives: Be able to define generative AI and key concepts and terms; learn about the types of generative AI; become familiar with the history and political economy of generative AI.
 
Readings:  Ethan Mollick (2024) Co-Intelligence: Living and Working with AI, Introduction
Podcast:  The Last Invention, Episode 1 (available on multiple podcast platforms)

Reference points:  
Key questions:  What are different views about the progress of AI, and why do they diverge?  What are the schools of thought about how AI will impact the labor market, and where do you land?  What are the impacts of the AI boom on the economy, and why do some analysts see it as a bubble?
Concepts: Jagged frontier, AI as normal technology, AGI / ASI / ANI, technological diffusion, S-curves, generative AI, generative adversarial networks, neural networks, model training, transformers, foundation model, parameters, tokens, context window

Week 2:  Alignment and prompting
 
Objectives:  Understand current approaches and concerns about AI alignment; learn and practice techniques for formulating and refining prompts including few-shot learning, chain of thought prompting, and more.

Readings:  Ethan Mollick (2024) Co-Intelligence: Living and Working with AI, Chapters 1 and 2
Leopold Aschenbrenner (2024), Situational Awareness: The Decade Ahead, Introduction

Reference points:
Key questions:  What are the elements of a good prompt?  How do computer scientists try to ensure AI models are aligned?  What does an alignment failure look like?
Concepts:  Context and constraints, chain-of-thought prompting, zero-shot and few-shot prompting, recursive self-improvement prompting, RLHF, superalignment, interpretability, safety card, system card
Exercise:  Discussing citywide solutions for lead in soil, using chain-of-thought prompting and assigning a persona.
 
Week 3:  Evaluating output

Objectives:  Learn and practice fact-checking strategies for dealing with hallucinations and errors as well as evaluating quality, reasoning, and bias in AI-generated output.
 
Readings:  Ethan Mollick (2024) Co-Intelligence: Living and Working with AI, Chapter 3
 
Reference points:
Key questions:  What is a "hallucination" and how are computer scientists addressing the problem?  How do professionals fact-check, and how are fact-checking skills applicable to evaluating AI output?  What other approaches are necessary to ensure reliability of output?  What is "quality" output?  By what criteria should we judge AI output?  What kinds of biases are people concerned about, and how can they be corrected?
Concepts:  lateral reading, editorial fact-checking; authority, currency, truth; relevance; bias; credibility; quality; open models vs. closed models
Exercises: Practice worksheet for identifying hallucinations using editorial fact-checking skills; practice worksheet for identifying AI slop; comparison of models exercise.

PART 2.  APPLICATIONS

Week 4:  Creating

Objective:  Learn strategies for using generative AI in ideation and creative work; practice creating a custom GPT

Readings:  Ethan Mollick (2024) Co-Intelligence: Living and Working with AI, Chapter 5
 Anil Doshi and Oliver Hauser (2024) “Generative AI enhances individual creativity but reduces the collective diversity of novel content.”  Science Advances (10)8.

Reference points:
Key questions:  What is creativity, and how have psychologists come to understand the creative process?  What is ideation, and how do design thinkers approach it?  What parts of the creative process might AI be useful in, and where is it not useful?
Concepts: creativity, ideation, creative process, rapid prototyping
Exercise:  Generating environmental solutions with AI using divergent thinking; using AI for implementation reality checks


Week 5:  Reading and understanding the literature
 
Objectives:  Explore how to use generative AI for literature review, creating structured summaries, and analyzing a body of research.

Readings:  Ethan Mollick (2024) Co-Intelligence: Living and Working with AI, Chapter 6

Reference points:
Key questions:  What are the qualities of a good literature review?  What does "AI-powered search" mean, technically speaking?  How do you know which search tool is best for your task?
Concepts: literature search, literature review, systematic review, topic modelling, Boolean search, deep search, deep research, lexical search, embedding model, natural language vs. keyword search, retrieval automated generation
Exercise:  Comparing models for deep research group presentations.
 
Week 6:  Writing and presentation
 
Objectives:  Understand how to use generative AI in improving organization and argument in various stages of research (proposal writing, research protocols, different types of final outputs, presentation materials), as well as common pitfalls.
 
Readings:  Ethan Mollick (2024) Co-Intelligence: Living and Working with AI, Chapter 7

Reference points:  
Key questions:  What makes a good presentation?  What are the qualities of a slide deck vs. an actual presentation?  What formulas for the structure of a good presentation are used in different contexts?  What rules of thumb make decks visually appealing?
Concepts:  SCQA framework, pitch deck, TED talk formula
Exercise:  Creating an annotated bibliography using NotebookLM; revising other people's papers using an LLM and comparing versions critically.

Week 7:  Teaching yourself new skills
 
Objectives: Learn about how people have used AI to improve how they learn; become introduced to AI agents; learn about fine tuning LLMs.
 
Readings:  Ethan Mollick (2024) Co-Intelligence: Living and Working with AI, Chapter 8
 
Reference points:
Key questions:  When and how can AI help you learn, and when can it impede your learning?
Concepts:  agentic AI, gamification
Exercise:  Tutoring prompt to study dimensions of war in Iran.

Week 8:  Image generation
 
Objective: Become familiar with the basic idea behind generative adversarial networks; learn about options for generative AI in image and video generation, and how these might be used in research.
 
Readings: Thomas Davidson (2024), “Start Generating: Harnessing Generative Artificial Intelligence for Sociological Research.”  Socius, vol. 10.
 Şerife Wong (2023), “The Origin of Clouds”, Logic magazine, issue 19. 
 
Reference points:
Key questions:  How does AI image generation work?  When and why might you want to use an AI-generated image, and when not?
Concepts:  GAN, diffusion models, JSON
Exercise:  Worksheet on exploring image generation tools, prompting for different purposes, and prompting with JSON.
 
Week 9:  Data analysis and visualization
 
Objectives: Learn more fundamentals about how non-generative AI and machine learning experts work with data.
 
Reference points:
  • Tools tested - Dify, Groq

Key questions:  When might you want to use an automated workflow to analyze data?  How do you verify AI-augmented data analysis?  What are the characteristics of a useful data visualization?  What are some tools for building interactive data visualizations?
Concepts:  API, API key, retrieval augmented generation
Exercise:  Analyzing environmental grants from New York state and building a classifier; answering questions about allocation of funds; creating an infographic.  Exercise producing visualizations of oil and gas well data in New York State.

Week 10:  Coding

Objectives: Examine how generative AI can help you write, review, and debug code for quantitative analysis and research, learn about the limitations of trying this without computer science fundamentals, and figure out how this might apply this to your own environmental research project interests.  Learn more about AI agents.
 
Readings:  Sarah Kessler (2024), Should You Still Learn to Code in an AI World?  The New York Times, Nov. 24.

Reference points:
Key questions:  What are the opportunities and limitations of using AI to do programming tasks, if you don't have computer science expertise?  What should you be careful about?
Concepts:  agentic workflows, functions, API calls
Exercise:  Creating a rainwater harvesting calculator.
 
Week 11:  Simulating and forecasting
 
Objectives: Identify ways to use generative AI in creating synthetic datasets and debates around this, and study and visualize different future scenarios.
Reading:  Choose one paper from https://www.climatechange.ai/papers?.
Reference points:
  • Tools tested: Vercel
Concepts:  world models, predictive intelligence, pivot tables
Exercise:  Analyzing sea level change, projecting sea level change into future climate change scenarios.
 
PART 3.  FUTURES
 
 
Week 12: Environmental implications of Generative AI
 
Objectives: Understand how analysts are thinking about the energy, climate, and water implications of computing and generative AI in particular, and identify key positions in the debates about how they might be addressed.
 
Readings:
Reference points:
Key questions:  What are the environmental implications of AI?  What technical and policy solutions are available to deal with them?
Concepts:  behind-the-meter, load growth, closed loop cooling
 
Week 13:  Understanding AI caution and optimism
 
Objectives: Understand the reasons scholars are cautious around how AI is developing, including bias, surveillance, how data is sourced, geopolitical considerations, and the AI workforce.  Understand the vision of the future that AI enthusiasts see.
 
Readings: 
Reference points:
Key questions:  What is a manifesto and what are people trying to do in this genre?  What are the types of bias people are concerned about with generative AI?  How are models trained, and how is this labor often exploitative?  What are the range of views about the social impacts of AI, and why do people hold such different perspectives?
Concepts:  bias (again); labor; industrial policy; universal basic income; token tax
Exercise:  AI policy research and presentations

Week 14:  Final presentations and reflection
 
Objective:  Share your personal projects and learnings with the classroom community.

So those are the topics... what will we actually learn?

We’re going to work on four things: concepts, skills, facts, and context.
 
Concepts are ideas that help us describe and explain things in the world: alignment, general purpose technologies, responsible AI, AI winter, public interest technology, and so on.  We will spend time learning about them, including who came up with them and why, when they might be useful, how to explain them to others, etc.
 
Skills are things your college education should be helping you develop:  finding information, collecting and analyzing data, communicating and storytelling, teamwork — in general, things that will be useful in life and in your next job.  This is a skills-focused course, focusing on specific skills related to generative AI use.
 
Facts are things like: When was the first neural network developed?  How many CO2 emissions are really associated with a ChatGPT query?  They are empirically verifiable pieces of information that tend to answer the who, what, when, and where.  In a world with Google as well as AI tools, you obviously don't need to know all the facts all the time.  But there are a few facts that are useful to memorize and have in your pocket for when you need them. 
 
Context is broader background knowledge (of which facts are just one part).  Context may be historical or spatial. 

What are the assignments?

Assignment: In-class assignments (50% of grade)
What is it?  There will be in-class assignments every week, often worksheets or lab exercises.  These will generally be graded on effort and completion. 
 
What is the purpose of it?  The purpose of these is for you to work through concepts using different examples, and for me to be able to frequently track your learning progress.
 
Assignment: Event reflection (10% of grade)
What is it?  A short paper you will write after attending an event either on generative AI or on data science / machine learning applied to environmental issues.  This event can be online or in person.
 
What is the purpose of it?  The main purpose is for you to see the kinds of discussions and debates happening among a community of people interested in genAI, and see the range of views and concerns people are having; to have a sense of how things are playing out right now.  The secondary purpose is for you to think critically about how events are designed and about public speaking, given that this is a part of many professional jobs.
 
Assignment: Portfolio piece (35% of grade)

What is it?  This will be a scaffolded assignment (meaning there are a few parts to it) that you work on over several weeks.  You will produce a piece of work that incorporates generative AI into an environmental topic — this could be developing a chatbot for people to learn about climate science, inventing a workflow, training your own model on an environmental dataset, or whatever.  You will have feedback from the professor to develop your concept.  It is designed to accommodate multiple levels of technical ambition.
 
What is the purpose of it?  The point is for you to be engaged in your own learning process through exploring a topic you’re interested in, and for you to have something you can show to potential employers or clients regarding your incorporation of AI into your work.
 
 
Assignment: Final reflection (5% of grade)

What is it?  This will be an in-class handwritten assignment that asks you questions to reflect on what you’ve learned and express your own ideas about this technology.

What is the purpose of it?  The point is to demonstrate your ability to think critically about how you will use this in your own work and life, and for you to assess your own learning.