Disruption often comes with a price. Over the last few weeks there’s been an extremely vocal online community writing inflammatory social media posts and misrepresentative blogs about the AI mapping experiment Ian and I have been undertaking. It’s like Lord of the Flies: The Cartography Years. Despite many more interested people (note – not necessarily positive, but open-minded) than not, there has to come a point where our side-project, done on our own time, becomes simply not worth it.

For decades, Cartography has faced numerous technological changes. The International Cartographic Association uses ‘Commissions’ as key instruments to explore how all manner of disruptions and developments might impact the domain. From work that people do, ICA brings forward ideas, and frameworks which help define cartography as it moves forward and reacts to change. Does it always work? No. Can you work outside of ICA and still offer critical ideas? Of course! Ian and I have been Chair and Co-Chair of the Commission on Map Design for the last 16 years (we flipped roles half-way through). It’s provided an avenue for work in the broad area of the design of maps that gives it some visibility beyond self-promotion. It engages an international body of professionals across the cartographic spectrum.

But we face a headache. At previous inflexion points of technological change Cartography has generally been slow to adopt or adapt. And then it’s almost too late. With the advent of AI, despite all its problems and concerns, we thought that a couple of fairly-seasoned experts jumping in to the mix to explore opportunities and difficulties might help us get ahead of the curve. We perhaps naively thought people would back the efforts we were making to at least try things out. We did so without an agenda. We wanted to explore, and to experiment. The simple idea of creating a map a day was born, using the sometimes serious, and sometimes frivolous ‘Day of…’ lists we always see. It was a convenient framework, and one neither of us had any particular reason to be biased about.

For several months at the start of the year it was difficult. The maps were tough to make. Generative AI was making some really bad mistakes, and web mapping was a challenge. But slowly AI Assistants have improved across the year and our ability to write prompts that do a better job of controlling the LLMs also improved. And just as we’ve been finding our feet along came the online backlash. No names, but the noise has become almost intolerable. We’re finding we have to defend very innocent motives from people IN the cartographic community who are making gross assumptions about our intent. Now, I get that folks will have opinions. But most are asserting those opinions purely based on the fact they “don’t like AI” (or similar), and if they can’t think of anything about the technology to moan about then they just don’t like our ‘tone’. That’s not really how science evolves. But since academic publishing is largely irrelevant any more, and codes of conduct make conferences an impossible forum for sensible debate, then blogs and social media is what we’re left with.

So, what to do… carry on regardless, with significantly decreased appetite, and either ignore or continue to bat away the haters (while trying to siphon out useful, mature critique). Or just end the experiment early. What is best for Ken and Ian? What is best for cartography? Are those two things the same? With upcoming conference presentations on the horizon we have commitments. And we’ve long-promised to report on what we find in our experiment to the wider community. But some already have us down as AI promoters so it’s largely pointless. They’re going to come to a conference presentation armed for a public take-down (except they won’t be able to do that due to the codes of conduct they were probably involved in writing in the first place). Ending a presentation by saying that voices from inside the community itself made it impossible to move forwards seems to me it would shine a very dim light on the way we act as a professional community. Let’s not wait until researchers have given us their conclusions, let’s shoot them down early, and often, instead! If no-one is interested in what we might have to say then the community has spoken, and we can just get on with doing other things. If there’s no interest in how AI might impact cartography based on a year’s worth of trying out what it has to offer then fine. That seems awfully narrow-minded to me but I could be wrong. What I do know is this experiment has eaten considerably into mine and Ian’s lives over the last 6+ months. Neither of us needed to do it. But having people so vocally shout-down our efforts has been surprising.

We haven’t made our minds up just yet because we believe what we’re trying to do has value for the wider community. We may even end up revealing that AI is a piece of crap for making maps. But that would be pre-empting even our own conversations and findings. It would be deeply frustrating to ditch this work just because of a few rather thoughtless souls but it does tend to be the case that the few always tends to spoil things for the many in most walks of life. Maybe this is just one of those occasions and we just have to shrug our shoulders and live with it.

So – if these blogs just end, or the posts on X, BlueSky, and LinkedIn fall silent you’ll know what happened. We just gave up, and gave in. If they continue then we’ll have persuaded each other that, for now at least, it’s worth trying to cut through the crap for a bigger end goal.

So…for now, here’s today’s map. It might be the last in our series. I don’t yet know.

A few years ago I made a wooden cheese board, and filled it with geographically appropriate cheeses that I served to a room full of geogeeks in London. It was a great, and tasty, success. I thought today I would try and prompt our AI map assistant to create a digital version.

Here’s the real-world version. Real wood. Real cheese. Taken with a real camera (with real sensors that captured the picture and converted them into glorious pixels)

Although far less tasty it’s also way more easily produced in a matter of minutes, costs significantly less (there was over $500 of cheese on the original), and is far easier to share. But even though what I’m sharing today is no more than pixels, much like the photograph above, yeah – I prefer the original. I like maps. But I like cheese more.

Original Prompt (ChatGPT)

A few years ago I hand made a cheeseboard in the style of a map of the United Kingdom. I then purchased different cheeses from different parts of the UK which I served on the cheeseboard to a room of geographers. It was a huge success. Could you create a digital replica of a cheeseboard of the UK with cheeses from across the map, positioned close to their origin of production. please label the cheeses (accurately and without spelling errors or hallucinations). Since cheese is most often paired with a fine wine please also show a glass or two of wine with the bottle and a playful label indicating fine taste in geographical wit.