AI and the natural world
It feels impossible to get away from now. The rise of AI is here, and we're seeing it everywhere we turn; especially if you're like me and are anywhere near tech conferences or tech companies. And even if you're not, ChatGPT is out in the wild and being used on everything from homework, to building resumes, to helping people write content courses.
I don't know if AI as we know it right now has a lot of staying power in all the places we're seeing it, but I think it's definitely going to hang on and stay in some capacity.
So, if it's somewhat inevitable, that got me thinking about how it can be used in nature. It does great at pattern recognition, pattern matching, aggregating data, and visualizing certain concepts. That made me think that it might be useful for mapping, conservation, or identification. So I've been looking around to see what AI models are currently working to help nature.
This is a growing list, but some of the applications are pretty cool!
Scivision: They're helping with wildlife classification and identification both on land and in the sea. Scivision is also being used in conjunction with mapping software to help scientists track how climate change impacts plant growth to help with food security in the future, and track coasts and coastal vegetation to better understand erosion and ocean rise.
iNaturalist: This is used for species identification by naturalists and citizen scientists. I love iNaturalist because it's such a wonderful way to connect with nature on your own or with others in events like the City Nature Challenge where people come together to take pictures and identify all types of wildlife, plants, and fungi.
Tracking iceberg loss: The Sentinel-1 satellite sends images of earth back to scientists. AI is helping scientists identify and map out ice bergs with more accuracy.
Air quality forecasting: NOAA is using AI to help with weather and air quality forecasting.
AIRES for SEEA: This will help scientists create ecosystem accounts that can be monitored and updated based on changes.