Is data collected from students always reliable when taking decisions on education?

from Diigo https://ift.tt/2zO9KHO
via IFTTT

A number of educational models like constructivist and experiential approaches to education have shown that performance (in the form of marks)  is not a necessarily reliable gauge for predicting that learning has taken place….and yet marks are often part of the data collected in order to determine where students go wrong. Furthermore, predictions made on unreliable or inconclusive collected data can do more harm than good.

One of the less promoted aspects of AI in education is the isolation of the learner from the environment and peers and yet robust AI systems should take these variables even more into account. This is perhaps the idea behind modern behavioural approaches to put the environment (of learning) back into the equation by designing ‘architectures’ that take into account the ‘physical, socio-cultural and administrative environments in which choices are framed’ (Knox et al, 2020).

In spite of this, the same arguments that go into the removal of the teacher from the learning equation are often voiced when talking about AIEd. There are still aspects of the learning process, often related to the community of learning, that are still absent from learning algorithms. These include notions surrounding the emotive aspect of learning. Will these aspects be truncated in favour of ‘cleaner solutions’?

References:

Knox, J., Williamson, B., & Bayne, S., (2020) Machine behaviourism:
future visions of ‘learnification’ and ‘datafication’ across humans and digital technologies, Learning, Media and Technology, 45:1, 31-45, DOI: 10.1080/17439884.2019.1623251