Teaching plant identification at a university in the age of artificial intelligence

Societal Impact Statement Society depends on experts able to correctly identify plants. This skill set is taught at university, classically using tools such as identification keys. The advent of artificial intelligence apps for identification, while benefiting society in many ways, poses a challenge for university education: Students may not see the need for learning skills beyond using an app. This may lead to a generation unable to verify and maintain artificial intelligence tools for plant identification. We suggest university teachers carefully adapt their courses so students are equipped with the skills to become experts proficient in the use of all tools. Summary Plant identification skills are essential for humanity. Universities train the next generation of experts who identify plants and maintain and develop the underlying taxonomic infrastructure. Apps using artificial intelligence can now identify plants with high accuracy and speed, record data and integrate it with additional information. These features make them highly attractive for a general public but also for the students we train at university. Classically used identification tools such as text‐based keys by comparison appear unnecessarily complex. We outline a risk in this development: the emergence of a generation unable to provide the very infrastructure on which artificial intelligence tools depend, and to verify their output independently. We suggest three guiding principles for critically engaging with identification apps at university: (1) Treat students as future experts. They are the ones who will build future infrastructure and need to be able to test one method with another. (2) Design exercises that cover all skill levels. Simple remembering and understanding, sufficient for the casual use of ID apps, falls short of the critical mindset we strive for in academia. (3) Emphasise primary data. Understanding that all botanical information, however much aggregated, is derived from physical specimens, is essential. Including ID apps thoughtfully in plant identification courses based on these principles can enable students to critically assess both their strengths and weaknesses. This will help ensure we continue to train experts proficient in the use of all tools for plant identification.