Table of Contents 1.

When you mention the term AI (artificial intelligence), chances are the person you’re speaking to will either glaze over or roll their eyes skyward, such is the effect that simply mentioning this emerging technology has on people. For some, AI heralds the next industrial revolution; for others, it is something akin to Skynet! But as with most divisive issues, the truth probably lies somewhere in the middle, and as AI advances, it will undoubtedly be employed in more and more applications, including in the wildlife sector.

It can’t have escaped many avid birders’ attention that the use of apps like the Cornell Lab’s Merlin app to help identify bird calls has increased over the last few years. The app uses machine learning developed by the Cornell Lab of Ornithology and relies on photos and sound recordings submitted to the eBird database by birders. This data is then archived by the Macaulay Library at the Cornell Lab and is funded and supported by the Macaulay Family Foundation, together with several other sources, enabling the app to be available free of charge. As more data is continually collected, the app’s ability to correctly identify birds will improve, along with the variety of species it recognises. Having technology like this in your pocket is not only impressive, but also very useful when you are unsure what you heard.

However, this kind of easily accessible technological ability will not always be appreciated by some, and while it is true that nothing beats experience and the development of field craft, having another opinion to confirm or deny your suspicions can be very helpful. For some time, there have been questions about the ethics of using data (images, text, or audio) as a teaching aid for AI. As a result, many organisations involved in storing and distributing such material are now offering an opt-out for those who submit to image libraries and publishing platforms. Undoubtedly, this issue will stay a hot topic within certain circles for many years to come.

AI bird identification has also been employed for image-based services, with major brands like Swarovski (one of the funders of the Merlin app) incorporating it into one of their binocular products. Many cloud-based photo storage applications also provide a basic level of identification when users upload images of birds. Where they source their dataset from is unclear, but with prominent names like ChatGPT involved, it is reasonable to suspect that they may be utilising their services.

For some time, nocmig has been becoming increasingly popular. This process involves recording the calls of migrating birds at night and then using software to produce spectrogram images that help identify various species. The data collected can provide more information on migration routes, timings, weather conditions, and many other factors. With AI now more widely available, you can purchase devices for use at home or at any other location you visit that record all species heard in a given area. The equipment then identifies the calls by species and also logs the number of calls heard together with other associated data. This information, combined with weather and location information, is then plotted on BirdWeather, a website openly accessible to anyone interested. This relatively new service is gradually expanding, with more locations being surveyed all the time.

The coming years of bird recording will be interesting and will provide birders with increasing amounts of data to better understand what is happening and identify what we can do to help out. But AI and technology won’t be for everyone, and there will always be a need for records, taken down with a pencil and paper, to be submitted from gardens and green spaces. Maybe we should look upon AI as another tool in the toolbox, rather than the new standard and source for future data.