The article discusses major ways in which recently emerged large language models such as ChatGPT can be used by modern scholars, teachers, and students. Although the capabilities of large language models are no doubt intriguing, these machinces are fundamentally just an instance of intellectual tools. Therefore, in the first part of the article the author examines the question of the relationship between intellectual artifacts and human cognition in general. The author considers the widely discussed hypothesis of extended cognition by Clark and Chalmers and concludes that it is highly problematic. Instead, it is argued that intellectual artefacts should be considered as part of the socio-cultural scaffolding of human cognition. From this perspective, the author discusses specific ways in which large language models can be used to help translate research papers into foreign languages, resolve technical problems of various types, and support effective education strategies. The author assumes that the responsible usage of large language models should benefit users themselves.
ChatGPT, large language models; scaffolded cognition; extended cognition; cognitive sciences.