This is the first in a series of articles that will deep dive into the exponential technologies that will have the biggest impact on not only learning and teaching but also leading schools in the 21st century. This post has been co-authored with Nick Kairinos and the Fountech team. Nick Kairinos is a global thought leader in AI through his work at Fountech, and was recently awarded the prestigious Outstanding Expert award at the CogX 2019 conference in London.

Definition of AI

Artificial Intelligence, Machine Learning, Deep Learning, Neural Networks, Supervised Learning, Knowledge Graphs … the list of buzzwords goes on. To set the context for the article, and to develop your understanding if you are a non-technical person, it is important to clearly define what Artificial Intelligence is in the first instance.

Artificial Intelligence (AI) can be loosely defined as software that mimics aspects of human behaviour, including amongst others learning, reasoning and knowledge representation. To a great extent, AI achieves this by employing algorithms to find patterns and extract useful information from (large amounts of) data.

Of the AI toolsets able to mimic human intelligence, at the core of is machine learning (ML). In contrast to conventional algorithms, ML algorithms are dynamic, meaning that their output constantly adjusts in response to the data they are exposed to. This makes ML algorithms applicable across many disciplines. ML algorithms are often grouped into three main categories;

  1. Supervised learning – algorithms which learn patterns and gain predictive ability by comparing their response to a set of inputs (also known as a ground truth).
  2. Unsupervised learning – algorithms that discover the structure of without knowing a ground truth, and often their result is the creation of clusters that group data with similar properties together.
  3. Reinforcement learning – algorithms that discover how to best respond to a set of inputs via an embedded reward-penalisation system.

An important category for education is that of natural language processing (NLP). These methods generally fall under supervised learning but because they deal with the understanding and usage of language by computers, they are often referred to separately.

NLP is further subdivided into two other categories, Natural Language Understanding (NLU) which deals with deriving human understandable context/meaning from digital text and Natural Language Generation (NLG) which deals with text composition into a human understandable form.


Opportunities in Education

For Learners

The possibilities for learners are endless and there are a number of companies already taking strides through their service and product offerings.

AI has already found its place in assessment and testing through a diagnosis of a student’s skill level by analysing their level through various tests, mapping it to developmental stages or curriculum requirements, differentiating learning responses and providing a personalised program of learning. One example of such a development is Soffos: The AI-Powered TutorBot™ whose aim is to democratise and personalise learning. This level of differentiated learning is often espoused in today’s classrooms but is extremely difficult and timeconsuming to achieve by teachers alone (Marshall, 2016).

With growing advancements in AI language understanding, the complex and timeconsuming task of hearing a student read, diagnosing their reading ability, identifying gaps in their development, addressing these gaps and recommending the complexity (level) of the next text to read, is certainly achievable.


For Teachers

The increasing burden and expectations on teachers have never been greater. In addition, the demand for teachers is projected to grow across the world as more countries move to provide both a basic and more advanced levels of education.

Goal 4 of UNESCO’s Sustainable Development Goals (SDG’s) is Quality Education with the goal being to “ensure inclusive and equitable quality education and promote lifelong learning opportunities for all”. Through AI, it is possible for teachers to augment their teaching. In this respect, AI solutions can be used to support teachers as opposed to replacing them, enabling teachers to spend more time with individual students, adding value at the individual level as opposed to feeling overwhelmed with planning, reporting and recording requirements.

Solutions that may help alleviate these pain points and utilise AI technology are:

  • Services that help teachers to quickly differentiate and personalise learning for all students, reducing their planning workload.
  • AI powered marking controlled by the teacher to enable quick feedback to the students which also provides an overview for teachers to know where they need to focus their teaching at an individual and class level.
  • AI generated reports based on ongoing formative assessments that are personalised and provided to students to give them a truly accurate indication of where they are at in their learning.


For School Leaders – system and in-school

A recent post by McKinsey suggested that AI has the “potential to help you lead with clarity, specificity and creativity.” While some may push back, particularly on the ‘creativity’ claim, there are many who would be keen to explore how AI might enable them to focus on what’s important, as opposed to what’s urgent.

With reams of data available at our fingertips, an interesting question for school leaders, and teachers to consider is ‘if you could collect actionable data on anything, what would that be?’ AI can help with both the collection and analysis of data. It can also be taught to explore a range of pathways rather than insist on a single destination; when leaders choose a specific action, the AI can offer an alternative set of choices and explore the possible implications of these. We all have cognitive biases and having AI supporting us to recognise these would be powerful in helping leaders make informed decisions.

Current possibilities for leaders include:

  • Coach as a chatbot – two options are LEADx and PocketConfidant.
  • AI Personal Assistant – which supports you in all aspects including identifying cognitive biases, taking minutes, scheduling meetings etc. There is an opportunity for a student and teacher version too.
  • AI Conversation Analysis – imagine being able to analyse the ‘quality’ of 1:1 and team conversations to leverage the most out of every conversation.
  • AI wellbeing pulse check – feedback on how students and staff are feeling and coping on a regular basis.
  • AI powered recruitment and retention – augmenting current practices to minimise bias and to give individualised professional learning support.


Implications of AI

Despite arguments made in the recent article, ‘Will AI take over Educational Leadership?’, it is our view that AI will always be used to augment rather than replace humans. School leaders and teachers play a very important role in the education system, to both achieving SDG 4’s goal of quality education for all and in achieving truly personalised learning.

This future state, however, will require significant changes to the current education system and we cannot forget that AI integration should always be about augmentation as opposed to replacement. This can only start through open consultation with the education sector, facilitated through the discussion, trialling and implementation of effective AI solutions for learners, teachers, leaders and systems. A sandbox approach that brings together schools and EdTech entrepreneurs has the potential to give us the greatest chance for success.

In a future piece we will be reviewing both the practical challenges that could hinder AI adoption by educators and the creative ways in which educators can overcome these obstacles and benefit from the full potential of AI.

Nick Burnett of LearnTech Lab
Nick Kairinos of Fountech

Leave a Reply