Author: Rob Farrow

Abstract: 

The application of artificial intelligence in AI is increasing, but there is a growing awareness of the profound ethical implications which are presently undertheorised. The emerging consensus is that there needs to be adequate transparency and explicability for the use of algorithms in education. This presentation provides an overview of AI in education (AIED) and characterises the requirement for explicability as a response to the ‘black box’ of machine learning.  It is argued that explicability should be understood as part of a wider socio-technical turn in AI, and that there is a strong case for implementing full transparency in AIED as a default position. Such transparency threatens to disrupt traditional pedagogical processes, and mediation strategies will be needed.  There are also instances where non-transparency may be justifiable and in these examples processes for auditing and governance.