Human-Centred Analytics/AI in Education

A heads-up that three collections will hit the streets this year focused on how we can design so that human needs and values are well and truly centre-stage in educational tools powered by data, analytics and AI. It will be good to have detailed ‘insider accounts’ from researcher/developers who are reflecting deeply on how values are baked into their design practices and the infrastructures they are building, and how different stakeholders can engage meaningfully in shaping design. I’m excited about the papers shaping up for these volumes, so watch out for their releases mid- and end-2019…

What’s the Problem with Learning Analytics? Journal of Learning Analytics. Invited Commentaries on Neil Selwyn’s LAK18 Keynote Talk, from Carolyn Rosé, Rebecca Ferguson, Paul Prinsloo & Alfred Essa (late 2019)

Diverse reflections on an article by Neil Selwyn, based on his provocative keynote address to the 2018 International Conference on Learning Analytics & Knowledge. [Replay the keynote]

Human-Centred Learning Analytics: Design Frameworks, Tools and Techniques. Journal of Learning Analytics, (Eds.)  Simon Buckingham Shum, Rebecca Ferguson, & Roberto Martinez-Maldonado. (mid-2019)

An important feature of the learning analytics community is our interest in the human factors in learning analytics systems. When learning analytics tools are used, their success or failure must be judged not only on technical criteria, but also by their adoption and effectiveness in schools, universities and workplaces. Often this is where the gulf between hype and reality becomes apparent. The complexities of embedding innovative technology in authentic contexts open a range of critical challenges for the field. The theme of the 2018 Learning Analytics and Knowledge conference (LAK18) Towards User-Centred Design — how stakeholders can, or must, be engaged in the design, deployment and assessment of learning analytics. LAK18 also held its first Participatory Design workshop. We invite contributions to this special section that explore these issues in more depth.

There are well-established research and design communities interested in human-centred design. The Human-Computer Interaction (HCI) community, for example, has worked hard to couple academic rigour with relevance in the fast-moving world of software design, evolving from an assemblage of disciplinary sciences towards practical design practices. Decades of work within communities under headings such as Participatory Design, User-Centred Design and Co-Design have led to many advances in theory, methodology and tools. ‘Human-centred’ can be defined at many levels, including the user interface, the impact on working practices, shifts in users’ power and control, and the values that are baked into the data models. The organisational obstacles to good user-centred design are well documented, since budget holders must be persuaded of the merits of investing money and effort in order to bring stakeholders into the design process. Most recently, of particular relevance to learning analytics, the user experience community has begun to engage with the specific challenges posed by interactive systems using machine learning. Building on decades of work in these communities, we need to apply and where necessary extend these approaches to the specific educational contexts in which we work. We therefore invite researchers and practitioners to submit theoretical, methodological, empirical and technical contributions including but not limited to:

Experiences deploying design processes that explicitly involve stakeholders (such as learners, educators, instructional designers, and leaders) in the co-design, co-creation or participatory design of analytics tools.

Evaluations of tools and techniques that have been effective in assessing how end-users make sense of, interact with, and act on analytics feedback.

Examples of how learning analytics systems can be made more transparent and accountable to different stakeholder groups.

Examples of how educational leaders can create the conditions for, or inadvertently undermine, human-centred learning analytics systems.

Examples of the benefits (and costs) that the adoption of human-centred design tools and techniques can bring to stakeholders.

Arguments/conceptual models/examples clarifying specific challenges of human-centred design for learning analytics, beyond those already well documented from other domains.


Learning Analytics and AI 2025: Politics, Pedagogy and Practices. British Journal of Educational Technology (50th Anniversary Special Issue), (Eds.) Simon Buckingham Shum & Rose Luckin. (late 2019)

While there is a growing chorus of justifiably cautionary voices about the dark sides of data, algorithms and machine intelligence when used uncritically in education, sometimes these are from commentators some distance from the ‘nuts and bolts’. This issue will provide accounts from insiders, all of whom have agreed to engage with the theme of “Politics, Pedagogy and Practices”, whose dynamics play out at many organisational scales:

Practices: We are seeking informed accounts of how these technologies come into being — the social and material practices of designing analytics and AI educational tools, and the related practices of educators and other stakeholders needed to deploy these tools.

Pedagogy: For some critics, analytics and AI equate to adopting a retrograde pedagogy from the industrial era. Any mention of quantification, or machine intelligence, evokes connotations of behaviourism or instructivism. Contributions to this issue will question such simplistic assumptions, illustrating a range of pedagogies and associated outcomes.

Politics:From international educational datasets gathered by governments and corporations, to personal apps, in a broad sense politics infuse any socio-technical infrastructure, because it mediates values and power. How do the researchers and developers of these tools frame their work in relation to concerns around values, ethics, and societal impact?

This issue will be written for a broad audience, introducing what is or soon will be possible, and describing strategies for taking into account data/algorithm/AI ethics. Written also for seasoned researchers, it will synthesise and clarify contemporary debates, providing a reference point for both teaching, teacher development and research.

Leave a Reply

You can use these XHTML tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>