If you started your vocation within the education sector some 30 years or more ago you will recall how the the World Wide Web steadily touched and transformed every service that your school, college or university provided. For instance, the language of the World Wide Web enabled the widespread adoption of learning management systems which transformed the design, delivery and management of teaching, learning and assessment across the education sector. Likewise, the language of cognitive services will bring about changes; but these changes are likely to be much more profound; touching and irrevocably reshaping every facet of the education sector. For example, how will teaching, learning and assessment be affected once we introduce another active agent into the setting; especially when teachers and institutions are increasingly managed by algorithms to support the student's journey through their studies?
This article explores three broad questions relating to the use of cognitive services by schools, colleges and universities; namely:
- How will institutions design, manage and deliver services to their respective communities if they leveraged the capabilities of machine learning or cognitive platforms?
- What additional value can be derived from the use of cognitive services?
- What should we be wary of when using cognitive services?
How would schools, colleges and universities institutions design, manage and deliver services to their respective communities if they leveraged the capabilities of machine learning or cognitive platforms?
When you are first introduced to cognitive services you soon come to the conclusion that many of the digital services within the education sector tend to be copies of processes, routines and habits that may become obsolete in the near future. This is especially true when you consider how traditional artefacts such as learning management systems, student information systems, library management systems or the myriad of education smartphone apps could be presented or used by a student or a teacher if they leveraged the capabilities of a machine learning or cognitive platform.
At a very simple and mechanistic level cloud based machine learning platforms or cognitive services are set to change the way students, teachers and support teams engage with software applications and web based services. For example, students simply have to converse with a voice enabled cognitive assistant to access every piece of information that pertains to their studies. If this is the case, what does it mean for the design and presentation of the services and applications that lie beyond the voice enabled service? It makes no sense for these services to persist in their current state.
This is not a hypothetical discussion. A growing number of educational institutions around the globe have either launched or are in the process of constructing cognitive services that will support the students, teachers and support teams on their campuses. Here are three examples from The Universidad Siglo 21 in Argentina, Deakin University in Australia and Bolton College in the UK.
The Universidad Siglo 21 in Argentina has implemented a virtual cognitive assistant which supports students with day-to-day enquiries about the university and life on the campus. The service also supports students with their enquiries about course content. The solution uses AIVO's agent bot. It's a great example of how cognitive services can be used to support students, teachers and support teams on the campus.
Genie is Deakin University's personalised digital assistant for students. The following video offers a short overview of the service; and it offers the viewer an insight regarding the development of these services.
The following promotional video highlights some of the services that can be delivered to students through Bolton College's Ada service.
Overtime, Siglo 21's chatbot, Deakin University's Genie, Bolton College's Ada service and others elsewhere will begin to respond to a wider range of day-to-day questions from students, teachers and support teams across all the domains that make up the student life cycle. They will also assist individuals perform routine tasks such as applying for courses, transferring students from one course to another, handing in coursework for marking, booking rooms around the campus and much more besides.
As well as supporting individuals with day-to-day operational tasks, routines and processes; these services will also support teaching, learning and assessment. For example, the Ada service is currently being used by teachers to assess students on Moodle, the College's learning management system enabling teachers to move away from the traditional multiple choice or drag and drop assessment activities that are typically associated with online tutorials. In the near future, the use of cognitive services such as natural language classification, natural language understanding and machine learning models will enable institutions to deploy services that will be capable of automatically marking and grading long form answers. These services will also be capable of offering real time feedback to students. The ILT Team at Bolton College is currently developing a solution which makes use of machine learning models to automatically assess student work. In the first instance, the solution will be used to review the work placement evaluations that are completed by students on the College's work experience app. Initial results look very promising.
What additional value can be derived from the use of cognitive services?
Your worldview changes when you hold a smartphone in the palm of your hand; knowing that your students are able to converse with it about their college and their studies. At the present moment in time these conversations start with day-to-day enquiries such as 'where is my next lesson', 'when is my next exam', 'where is my work placement' and they gradually move to towards conversations that may start with 'what are my options once I graduate from my course' or 'help me complete my evaluation for the work placement that I have just completed'. The work that is being undertaken on a number of campuses around the globe bodes well for a future where cognitive services will offer additional value to students.
The following diagram highlights were cognitive services could add value to schools, colleges or universities and to the students who attend these institutions.  The use of machine learning within the education sector is based on a simple premise; namely that education institutions are essentially information-processing entities. If this is the case, then student and institutional data can be leveraged by cognitive service platforms to inform the myriad of algorithms that will come to inform the behaviour of students, teachers and support teams on the campus.
As algorithms begin to play a greater role on the campus they will invariably augment the capabilities of teachers and support teams; and by doing so they will enable the institution to be better placed at providing on-demand services for students and thereby add value to every aspect of the student life cycle. Please note that data is defined in its broadest sense; encompassing structured and unstructured data. Unstructured data includes text, images, the content library of a learning management system, every electronic document on the campus, social media posts and even sentiment.
What should we be wary of when using cognitive services?
I wrote an article in April 2013 entitled Technology and the Self which explored the relationship between technology and ourselves. If technology is playing such a vital role in shaping who we are, what does it mean for schools, colleges and universities when it comes to adopting and using technology? The work of Prof. Sherry Turkle extends the theme by stating that we need to move away from the commonly held view that computers are just tools that can be used to support teaching and learning (e.g to enable teachers to present information to the class, to support the creation and distribution of learning materials, to record attendance and grades; and for students to respond to set tasks and assignments given by their teachers) to a position were we are asking about what computers and networked devices and services are doing to us.  The introduction of networked devices and services in our schools, colleges and universities has amplified and sharpened the debate and interplay between the individual identity of students and the educational institutions that serve them. The following video is from the Big Thinkers series and it showcases the work of Prof. Sherry Turkle. 
In the following video Sherry Turkle advises teachers and educational leaders to be wary of creating conversational AI services that suggest empathy with students.
However, as conversational services mature they will respond with words of encouragement or praise to encourage positive action or behaviour in the student. They will also nudge or guide students towards desired actions or behaviours that would support their studies. For example, these conversational services could express words of praise when a student has achieved the desired average grade on the course; and having realised a higher grade profile, they may nudge or suggest opportunities for further or higher study; or another course of action that is seen to be desirable by teachers and support teams.
One of the core reasons for using machine learning is that it enables teachers and support teams to manage the increasing volume of structured and unstructured data that pervades the campus. The algorithms that reside on an institution's cognitive platform also enable teachers and support teams to make informed decisions and more importantly, timely and better decisions. Nevertheless, are students, teachers and campus administrators happy to be governed and managed by algorithms which shape what they see and how they perceive the immediate world around them; and are we happy about the bearing that these algorithms have on the decisions that we make or rather the decisions that they make on our behalf?
In another context, adaptive learning environments will subtly alter the way colleagues teach and assess. Teachers who embrace the potential benefits of adaptive learning environments will have to get used to the fact that another agent; namely the adaptive learning environment, will begin to determine the content and assessment activities to present to their students during the online component of their courses. Teachers will need to see evidence of improved student outcomes, improved student retention and higher student satisfaction ratings for their courses if they are to accept the growing use of adaptive learning environments in their institutions.
If courses are delivered partly or wholly online what will the reaction be from students who learn that they are being taught by hundreds and thousands of algorithms or by an adaptive learning environment with little or no input from a qualified teacher? How will the teaching profession react to the increasing use of adaptive learning environments? Will teachers feel that they are being displaced by adaptive learning environments? And how will parents react when they see their children being taught by an adaptive learning environment? These questions, and many more will need to be addressed if adaptive learning environments are to become the norm within the education sector.
Algorithms are increasingly playing an important role in the way we select students onto a course; the way we shape and deliver teaching, learning and assessment; the choices that students make during the course of their studies and more. The nature of teaching and the services that are offered to support students of all ages means that explainable AI is set to become the norm. It will be unacceptable to deploy an algorithm on the campus whose decisions cannot be readily deciphered by a teacher or a member of the support team. However, an increase in general transparency could lead to a lessening in the capabilities of the algorithms that are used to support students.