Evaluating the accuracy of the Seek app for conifer identification: A baseline for future educational use
Year: 2025
Type: article
Abstract: In recent years, mobile applications, including artificial intelligence‐based ones such as Seek by iNaturalist, have become essential educational tools, particularly for science and outdoor learning. Therefore, we decided to work with this mobile app and conduct a pilot investigation of its reliability in selected conifer species identification. We focused on conifers because they have green leaves all year, and they are found almost everywhere—they are widely distributed in cities with many representatives of different species. We tested 40 conifers from a local botanical garden, and our study assessed Seek's accuracy in identifying them, showing an average accuracy rate of 38.75%. The selection of conifers included both common species native to Czechia and rarer species from common genera, as well as non‐native conifers frequently cultivated in Czech parks. Common species or species with distinct morphological features (e.g., Larix decidua , Picea abies , Thuja occidentalis , Cryptomeria japonica , or Metasequoia glyptostroboides ) were identified more reliably than lesser‐known species (such as Larix kaempferi or Pinus cembra ). Although initially focused on North American flora, Seek is also useful in European contexts. Despite its potential, also for future use in education, users should remain aware of its limitations, particularly in identifying less common species. Nevertheless, all basic taxa were identified to some extent, providing valuable support, for example, for research‐based outdoor education.
Source: Natural sciences education
DOI: 10.1002/nse2.70030
Institution: Palacký University Olomouc
Language: English
Subfield: Social Psychology
Field: Psychology
Domain: Social Sciences
SDG: Life on land
Open Access status: hybrid
Estimated APC paid: $1,580
OpenAlex ID: W4414627443