Adaptive learning is promoted as a
form of personal learning, with opportunities to tailor learning to students’
understanding. It is perhaps most useful when students have to learn
foundational principles or practice essential skills. Adaptive learning resources take a large amount of time to develop. The learning designer requires a good understanding and ability to unpack the tacit dimensions of learning. Adaptive learning is good for helping students learn the ‘canon’ required – especially foundational theory and discrete skill sets (e.g. basic maths) which have one ‘right answer or recommended way to solve.
Adaptive learning has been around for some time, for example, 'programmed learning' 30 years ago was available mostly through text-based resources and I remember testing out very basic computerised versions. Basically, they were text books with small (usually multiple choice) quizzes and the results from these quizzes, directed you to another part of the text book. The approach was based on behaviourist theories which emphasised scaffolded learning.
Currently, adaptive learning is again and a response to standardised learning promoted in many countries - see previous blog for overview and return of adaptive learning to the list of currently recommended pedagogical approaches. in the 'new' iteration, adaptive learning is defined as the ability of a learning resource to adapt to learners' performance. Edsurge reports advantages and particularly for online learning. Educause article also supports adaptive learning as a means to achieve successful learning. There is a good article by Kerr, P (2015) on the topic providing definitions for '‘individualization, differentiation, personalization’ with adaptive learning being the technology rather than the approach.
With the advent of AI, adaptive learning may be one approach to achieving economies of scale with blended / online learning. Education dive, lists many adaptive learning platforms, with many being publishers of text books and other forms of education resources. Forbes reports an upsurge in adaptive learning platforms with a more up to date list from tech advocate.
Commercial offerings include smartsparrow (free for up to 5 learners and up to 100 learners cost US$15 each); dreambox; knewton; and adaptemy.
Open source platforms include a Harvard and Microsoft collaboration, alosilaps; grapple; and sagefy.
However, there are always other
factors to consider. One being the lack of learner choice as the algorithm
directs learners on pathways which the learner may not have envisaged going.
As proposed by Siemens, adaptivelearning may be constraining. Another approach is to perhaps offer learners
sufficient support to understand the outputs of learning analytics and then for
them to work out ways to address the data - see slideshare for 2015 presentation from Siemens, Gasevic and Baker.. The learner has to learn the skills
to interpret and act on learning analytics, instead of being taken, without
understanding why, down pre-programmed pathways laid out through adaptive
learning platforms.
The other current
challenge with regards to deploying adaptive learning with vocational
education, is the prevalence and reliance on text based approaches. The advent
of VR may provide an opportunity to move beyond adaptive learning based on text
based responses but for the present, multiliteracies and multimodalities are
not commonly included in algorithms for adaptive learning.
So, the main learning from undertaking the exercise of exploring adaptive learning, is caution. Adaptive learning offers advantages but can be constraining and may lead to learners being forced down the linear syllabi path via behavourist approaches. Balance needs to be sought, to provide learners with greater agency as to how learning pathways may be completed.