AI Naturalists Might Hold the Key to Unlocking Biodiversity Data in Social Media Imagery

Abstract

Summary The increasing availability of digital images, coupled with sophisticated artificial intelligence (AI) techniques for image classification, presents an exciting opportunity for biodiversity researchers to create new datasets of species observations. We investigated whether an AI plant species classifier could extract previously unexploited biodiversity data from social media photos (Flickr). We found over 60,000 geolocated images tagged with the keyword “flower” across an urban and rural location in the UK and classified these using AI, reviewing these identifications and assessing the representativeness of images. Images were predominantly biodiversity focused, showing single species. Non-native garden plants dominated, particularly in the urban setting. The AI classifier performed best when photos were focused on single native species in wild situations but also performed well at higher taxonomic levels (genus and family), even when images substantially deviated from this. We present a checklist of questions that should be considered when undertaking a similar analysis.

Passive citizen science: The role of social media in wildlife observations

“…These same patterns were mirrored at the class level with the highest number of returns for Flickr being Magnoliopsida , a class of flowering plants, and the second highest was Aves . These results also support findings presented in [ 11 ] where results showed that Flickr is rich in plant-related data mainly dominated by non-native flowering plants. However, that research focused only on plants and does not perform comparison between Flickr and citizen science data portals.…”

“…The use of iEcology approaches is increasing as it provides low-cost and fast data collection, pattern identification, and visualisation of nature-related data. In particular, the study of [ 11 ] investigates whether a plant species image classifier can be used to extract relevant plant observations from Flickr, using a general search term of ‘flower’. Analyses showed that automated methods have the potential to help identify wildlife-related imagery data on social media, especially when photos were focused on single native species in rural situations or when classification was performed at genus or class level.…”

“…This previous research was conducted on a narrow range of species (between two and four). Another similar research by [ 11 ] investigate whether an image classifier for identifying plants could facilitate the discovery of unexploited biodiversity data from Flickr. However, this approach is focused purely on species occurrence on Flickr and thus does not provide a clear evaluation of the role of social sites observations compared to more traditional approaches.…”

“…Social network websites such as Flickr, Twitter, and Facebook have built a network of more than 2 billion users worldwide, generating millions of messages daily that are easily accessible, and reflect the observed reality of a quarter of the human population [9]. Social media websites have therefore emerged as an informal real-time information source that can contribute to the detection of trends and early warnings in critical fields such as ecological change, environmental problems, and shifts in ecosystems [9][10][11]. We define the use of social media that are unconnected to any particular citizen science program, but represent an unexploited source of valuable ecological data as passive citizen science.…”

classecol: Classifiers to understand public opinions of nature

“…In recent years, the use, and diversity of uses, of social media analysis across the environmental sciences has rapidly increased (Ghermandi & Sinclair, 2019). Social media has been used to develop species distribution models (August, Pescott, Joly, & Bonnet, 2020), measure aesthetic and recreational ecosystem services (Van Zanten et al, 2016;Graham & Eigenbrod, 2019), track illegal wildlife trade (Di Minin, Fink, Tenkanen, & Hiippala, 2018), and determine the role of wildlife in nature-based tourism (Hausmann et al, 2017). The abundance and availability of data on these platforms -many now 15 years old, open the door for more research.…”