ABSTRACT
AI’s rapid development promises disruption and reshaping of learning with widely divergent projected outcomes: From fears of plagiarism and cognitive offloading, whereby students lose out on learning opportunities by delegating thinking and learning to AI, to an endless learning resource, partner, and even interlocutor in learning processes, generator of personalised learning content and experiences, etc. Additionally, AI is often presented as potentially enhancing education by improving accessibility, optimising administrative tasks, and preparing students for a technology-driven workforce. However, the optimist camp is also characterised by deeper antinomies, including questions as to what and whose values are to guide education, what the effects of data-driven predictive processes in education are going to be, and how (and what) content is to be retrieved. As such, the main AI crossroads in education appear to be between autonomy and integration and efficiency versus diversity. In the latter pair, efficiency pertains to datafied optimisation of learning processes with quantity and speed as a priority; diversity, meanwhile, encompasses openness to a broader set of values (including culturally and situationally specific ones) and problematisation of knowledge that may not allow for an optimised learning process but could retain it grounded in specific experiences and concerns.
