How to Rewild

How AI Will Change Ecology

Home / Magazine

How AI Will Change Ecology

Magazine

Contents

In February 2024, I witnessed something remarkable. On an obscure video development page, a collection of clips were released which heralded the dawn of a new era in ecology. 

In one, the camera flies through a California gold rush town, as horses flicker occasionally into and out of existence. In another, we watch from a disembodied drone, as waves crash endlessly onto golden cliffs (below).

All of the clips on this site were hallucinated by a machine – generated by artificial intelligence. The next step in a technology which is rapidly replacing jobs and opening up new frontiers. In a decade, AI imagery has gone from blurry blob renderings to flawless HD video of entire ecosystems. In another decade, no doubt we’ll be walking through those ecosystems in VR, courtesy of headsets like Apple’s Vision Pro.

Using AI Today

Is this scary? Thrilling? One thing’s for certain – failing to work with AI will mean falling behind.

Life to Land wouldn’t be possible without some AI assistance. Not, of course, in writing or researching the content – the potential for errors are too significant, and the reputational risk goes without saying. AI has a tendency to ‘hallucinate’, which is both a blessing and a curse – in one of the new videos created by the algorithm, endless puppies spawn from an apparent wormhole in the spacetime continuum. In writing, the same problem is true – if asked a question without an easy solution, the machine may hallucinate facts to come up with a better answer.

But it’s getting smarter by the day. Every time humans successfully trick the machine into hallucinating, and post publicly about it, the data gets fed back into the algorithm and the same feat gets harder to replicate. My own attempts to fool Open AI’s chatbot, ChatGPT, came up fruitless. Which is a little frustrating, as I’ve found that it can hallucinate frequently when using it for coding. Working with the chatbot is a lot like working with a real coder, only it’s free, available at any hour of the day, and an expert in every coding language you throw at it. Many challenging parts of the website, like creating To Do lists, would have been next to impossible for me (inexperienced with code) without the help of the algorithm.

Consultations
An AI-generated image of a British ecosystem (this looks inspired by the New Forest).

Alternate Timelines

Outside of the computing realm, the implications of AI are perhaps more terrifying. Back in 2018, I wrote a dystopian novel about the impact of neural implants. In 2024, Elon Musk’s Neuralink company reached the human trial stage and that world edged a step closer. With a direct connection into the brain, coinciding with AI-generated worlds, is there a danger that the real world becomes devalued? That people fall into a Matrix-like alternate reality and become disconnected from the ecological devastation around them?

Or could this tech be used to sell nature recovery? Already, we can input an image of a degraded ecosystem, or a sketch of a habitat as we imagine it, and generate a near-photorealistic render of the future. Of a healthier ecosystem, with vibrant colours, mature trees and wild animals – 10 years, 20 years, 50 years on. Video renders are maybe a year or two away from widespread public accessibility, and it sounds like stepping into this imagined future in VR could be on the horizon, too. This kind of realistic render would be far more compelling to a wavering landowner than facts, figures and sketches on paper.

First Look Consultation
An AI-generated image of a possible past British ecosystem - note, however, the American Bison (despite feeding the prompt 'European Bison')

Walking Through the Future

When you’ve walked through the future of your project, you’ll better-recognise its value in the present. And better-understand the interventions that you need to make, to create further biodiversity and productivity. These AI worlds could also be a gateway into our past – in fact, they already are.

The most remarkable clip for me in that smorgasbord of hallucinations was the helicopter journey through a California gold rush town (below). It showed that, drawing on historical photos, period films, maps and sketches, an AI may be able to recreate past ecosystems. Not just in the form of images, but in videos and, potentially, VR experiences that bring what we’ve lost from our landscape back to life. Rather than describing the activity of beavers and bison, we’ll be able to witness them in our own back garden, in an imagined past. Like a curated acid trip with educational benefits, that could push forward the cause of nature recovery.

Stepping into the Past

I’ve already seen the benefits of inhabiting painstakingly-recreated ecosystems from our past. The game Red Dead Redemption II (above) is a painterly approach at restoring the lost history of America’s Wild West, when the frontier still remained wild and unexplored in many places. The map is enormous – it takes perhaps an hour on horseback to ride from one end to another, through desert, dense coniferous woodland, swamp and prairie. Along the way, you can spot wild animals native to the region through binoculars, and collect pelts, feathers and shells. It’s a wonderful place to explore, and it would be captivating to see a similarly-realised vision of Stone Age Britain – to see what we once had, and may yet have again.

So, I’m positive about the potential of technology to capture people’s imagination. To foster a passion for ecological restoration and a return of wildness. As the expression goes, sometimes, we just need to touch grass, and I don’t think there will ever be a real substitute for that. At least, I hope not. Because if it’s ever possible to fully escape into a hallucination, then the machines will have triumphed over our humanity.

Further Reading

If you’re curious about the future development of your project, why not take a journey through the 7 Stages of Succession?

Contents

Discover Another Article...