What is possible with AI in GIS Education

GeriMiller · ‎06-01-2025

Purpose of Blog

AI has had a profound impact on how we do our work. This blog aims at elaborating on what can be done with AI in GIS, including Education examples, and where to go as next steps.

It is a series of 4 blogs:

When we think about AI in GIS, there are two approaches, GeoAI and AI Assistants. Importantly, there is an underlining framework of data and technology knowledge that is necessary, before leveraging these capabilities:

GeoAI – tools that automate processes and workflows, next steps to advance GIScience.
AI Assistants/ Generative AI – creating better experiences when using GIS tools, boosting productivity, making people more productive at using GIS.

Before you get started (framework/items to think about)

Before delving into some of these AI approaches, options and tools, there are important considerations to think about and keep in mind – a framework upon which we can build.

Having a clear problem to solve - the problems we are trying to solve should drive the analysis (and how we use these tools), not the other way around. Understanding what you are doing and why you are doing it (or what/why the AI is doing), is crucial. It is important to ensure that AI is the right choice to solve the problem.
Understanding foundational concepts - to leverage GeoAI or AI assistants, it is very important to understand the relevant foundational concepts before using the tools. For example, to leverage the deep learning tools with Imagery, which can deliver great results - one still needs to understand the fundamentals of Remote Sensing: data resolution, electromagnetic spectrum, temporal frequency, simple statistical classifications, etc. - i.e. some of the needed basics, as well as understand deep learning model architectures, parameters to use, etc.
Knowing the data – one must know the data they are working with, to use the tools successfully. The concept of “garbage in-garbage out” applies heavily to AI. In addition, it is important to learn what kind of data that a particular AI tool you want to use can work on. Each of the AI tools and pre-trained models come with metadata, guides, and transparency cards. It is important to follow the guidelines and data types to ensure success.
Having a good understanding of ArcGIS – once the foundational concepts are clear, one would use tools to solve a problem. Understanding the tools themselves, and how they work, is important.
Always keeping accuracy and ethics in mind – AI can introduce inaccuracies and lead toward a wrong path. One needs to critically evaluate results and have a degree of confidence in the output. One needs to be able to analyze where and when the model didn’t work, or why it may be failing.

GeoAI

GeoAI – we defined GeoAI as advancing the Science of GIS, with AI models, tools and techniques - to automate data extraction at scale and uncover valuable insights faster than ever. The first blog in our series highlighted several different examples of GeoAI, which were being used to perform predictive analytics or extract data. For example, in the past, it was a laborious effort to manually identify features in imagery, but GeoAI methods allow us to train deep learning models to extract these features. There are many such examples of improved workflows.
Education patterns and examples – below are options of how GeoAI techniques can be leveraged.
- Within a class (section/module/assignment)
- Entire course focused on AI
- Entire program on AI
- DIY (self learning, informal education)
- Education use cases – below are a handful of examples, there are many others:
  - University of Pittsburgh – example of section/module/assignment using AI as part of an introductory course
    - Leveraged deep learning frameworks to help teach modern GIS skills in introductory GIS course
    - Used a simplified deep learning installer for ArcGIS, which streamlines user experience
    - Inspired students to leverage learned techniques in other courses (Electives or Capstones), students hired by employers.
    - Read story here.
  - Johns Hopkins University – example of entire course, titled “Artificial Intelligence and Machine Learning in Geospatial Technology”. Topics covered:
    - Start, install setup and stop a virtual machine with GeoAI software components
    - Manipulate spatial data and perform numerical operations using Python
    - Visualize spatial data in various forms
    - Perform spatial data clustering using multiple algorithms, such as k-means and DBSCAN
    - Perform regression and classification on spatial data using multiple algorithms, such as GWR and XGBoost
    - Perform spatial object recognition using neural network
    - See sample Syllabi/Topics attached.
  - University of Florida – example of entire certificate/program/department
    - Certificate titled “Geographic Artificial Intelligence and Big Data”
    - More details here.
  - University of Buffalo – example of entire certificate/program/department
    - Department titled “Department of AI and Society”. More details here.

AI Assistants

AI Assistants – we defined AI assistants as tools which create more natural and intuitive experiences in ArcGIS, to empower GIS users and boost productivity. There are two options, broadly speaking:
- AI assistants (Esri)
- AI assistants (general, i.e. ChatGPT)
Education patterns and examples – below are a handful of examples, there are many others. When it comes to AI assistants, the “art of asking meaningful questions” is a crucial concept to teach/discuss.
- Help with learning concepts (personalized learning assistant)
- Get ideas (on projects, workflows, methods, presentations)
- Troubleshoot when stuck
- Verify results (of non-AI work)
- Education Examples:
  - Columbia University
    - Data Analysis course, encourages students to leverage ChatGPT if they don’t understand programming or data analysis concepts, AI can help learn in an easy-to-understand way.
  - Johns Hopkins University
    - Encouragement for any student to leverage AI for troubleshooting steps, or ideas for projects.
    - Example of student with ADHD who is struggling with reading comprehension especially with long texts – using AI to help brainstorm and breakdown how theoretical concepts tie into material that is taught. AI generates 1–2-page brief on subject matter, then via conversation, it walks through understanding the concepts.
  - Massachusetts Maritime Academy
    - Use in Applied GIS course, part of Marine Science, Safety, and Environmental Protection (MSSEP) course offerings, where students get introduced to Python
    - Purposefully unstructured approach - using AI to learn about different concepts. Encouragement to ask questions, help develop basic code, build scripts, check results, identify a task for their final project to use generative AI

Resources, where to go for help

Please check the next blogs in the series, “Teaching with AI”.

There are a lot of resources that can be used to get started. Please take a look at the below, and comment on what else could be helpful.

Getting started
Learn Tutorials and Esri Academy (search for AI, GeoAI, AI Assistants, example tutorial series)
Documentation and sample packages

If you need to speak to a person, please reach out:

United States
- Your Account Manager
- highered@esri.com
Outside United States
- Your Esri distributor