Generative AI in the geospatial sector

As a consultant and geospatial professional, I'm always looking to embrace new ways of working, to build skills and knowledge, and deliver the best service for the customer. This involves keeping up-to-date with product development and innovation, seeing what we can put to good use.

In the wider field of Artificial Intelligence (AI) and specifically Generative AI, there is so much for geospatial teams to understand and benefit from. I wanted to share some practical ways to get started, taken both from the experiences of our customers and from our own journey over the last few years.

What is GenAI?

Generative Artificial Intelligence (GenAI) has become one of the most talked-about technology trends, due to its ability to create new content including text, images, videos, and even geospatial data. It is one of the most talked about technology trends of recent years, with frequent new discoveries and applications.

Behind the hype, businesses are looking for and finding real applications of GenAI that is transforming how their teams work. By speaking with our government and industry customers and partners, we learned how they are applying GenAI in their organisations, their successes, and challenges.

Where to start?

So, how can GenAI help you, your team, and your organisation work smarter in your day-to-day work, improving efficiency and making a positive difference for you and your customers?

You might not know where to start; perhaps you think you don’t have the right expertise in place; or maybe your current technology lacks the capability. But there are opportunities available right now to achieve some immediate impact and help shift perceptions of GenAI within your teams.

Some real examples from our customers and from within OS where GenAI has had a positive benefit:

1. Build that killer web-map feature

A lot of our customers talk about the need to create engaging content for their stakeholders, and there are a wide variety of ways to do this on the web through maps. There are solutions that can get you up and running quickly, but if you want something more bespoke, you often have to code new features.

This is where GenAI can generate code that supports both open-source and commercial geospatial languages and libraries.

So, if you want to find that competitive edge by creating the feature that will wow your customers, use GenAI to get you up and running and delivering those features faster.

2. Present your work with thorough research

All geospatial teams at some point need to present their work, perhaps as a presentation or a report. And often this needs to be backed up by research into the use cases and subjects being considered.

Where web search tools can help find relevant information to support a task, GenAI can do so in a more conversational and evidenced way. It enables you to dig deeper, more quickly into the topic you are researching, like talking to a knowledgeable colleague.

This form of GenAI doesn’t require any experience in coding or data science; if anything, it requires a more creative set of skills around how to ask the right question to get the best results. With a bit of trial and error you’ll find that you can learn more in less time and reference your sources with little effort, leaving you more time to think about what you want to say.

3. Create and validate your geospatial data

GenAI is now available in a range of geospatial tools and enables you to train models to extract features, called Automated Feature Extraction (AFE). In OS, we fly planes to collect aerial photography for Great Britain every year, then identify specific features or changes from that photography to update our maps. For example, AFE can be taught to identify and calculate percentage values for natural land cover types – arable/grazing land, non-coniferous trees, mixed trees, and more – and the time saved can be significant, particularly when you consider doing this for large areas.

I would advise caution on any vendor statements as to the quality of their AFE models and thoroughly test them for your own use cases. AFE is still in its early days in an enterprise context, so it will suit some use cases better than others.

So what should you do?

GenAI can help geospatial teams work smarter today, with even greater potential in the years to come. But there is reticence with some of our customers, around the time they have available to test and play with GenAI, and the ability of their current tools to support GenAI.

What I suggest is that if you can prove, by some means, that GenAI has potential for your team, you can build support in your wider organisation. So try and follow these three principles to start your GenAI journey and help transform how you, your team, and organisation works.

1. Learn as a team. Get your team together and workshop how GenAI can help you replace or speed up tasks. Sam Altman, CEO of OpenAI, discussed thinking of tasks that take you 5 seconds, 5 minutes, 5 hours, or 5 days. Try that as a structure and you’ll find the ideas come quickly. Make GenAI play a part in making things better for your team.
2. Stick with it. Start using GenAI now for simple things, and you'll naturally see new features appear over time. The results might not be perfect but will only get better. Be part of the conversation rather than on the outside.
3. Share your stories. Share your experiences with your wider organisation and please get in touch with us to share your successes and challenges. Working together as a geospatial community will help all of us get the most from this exciting new technology.

As a final thought, there are lots of questions around how to use GenAI compliantly, ethically, and sustainably. Your organisation may have policies already in place, such as responsible use of AI, or may just be starting to get to grips with GenAI.

If you're not sure whether you should be using GenAI for a task, approach your compliance or security teams, and ask for help.

And please don’t hesitate to use LinkedIn to contact myself or Harry Searles, a geospatial consultant for OS, to discuss how you and your teams are using GenAI, or if you want to learn more about how you could get started – we'd be delighted to hear from you.

(This article was written by Tim Chilton, and supported by research undertaken by Harry Searles - click their names to connect through their LinkedIn profiles)