At the Designing Online Learning for the Future workshop (#OUDforL) our first keynote speaker, Donald Clark, shared his levels of Artificial Intelligence (AI). They are:


This relates to the software embedded in technology which, through the use of algorithms, enable devices such as smart phones or virtual reality headsets to interpret information from tiltmeters and clinometers.


We’ve all used it…  it’s called Google and runs on an algorithm.

Here at the Open University, the Supportive Automated Feedback for Short Essay Answers (SAFeSEA) project developed ‘openEssayist’. ‘openEssayist’ is a content free tool which uses natural language processing techniques to offer automated feedback of students’ draft essays. There’s even an artificial intelligence assistant, Amy, to schedule your meetings when you copy her/it in on emails. Sadly there’s a waiting list!

Assistive AI can also validate the identity of users using facial recognition and/or the analysis of behaviour such as keystrokes eye movement. Apparently Pearson VUE use this technology in their test centres.


There are already learning platforms claiming to offer customised learning for individual users. Essentially, a diagnostic and algorithm are used to determine which content from a course should be presented to the user. Examples include training for Microsoft Excel.


This is where we start to see technology becoming automated, utilising analytics data to adjust the course in real time to create a networked ‘path’ through different learning objects to suit the individual learner. The AI is also capable of detecting where students make mistakes, e.g. in a mathematical calculation, and is able to show how to correct the error in the working.


This comes in the form of immersive simulations; the use of analytics to diagnose and recommend actions; and performance intelligence. The latter is particularly interesting as it has the potential to do things humans can’t, e.g. more accurately predicting where crimes might happen or human behaviour. In the case of Amazon, they’re working on predicting what you’re likely to buy, before you know yourself, and begin dispatching it before you’ve even ordered it.

Google’s AI recently beat the European champion of GO (the board game). To do this, it needed to ‘learn’ the game and anticipate its opponent’s moves (rather than rely on a predetermined algorithm). More practical, although admittedly futuristic, applications of AI include Tesla’s self-driving car, which reportedly has the ability to be ‘summonsed’ from the garage to the user’s driveway. Uber probably have a close eye on this.

How will AI influence both our approaches to designing learning and the learning experiences we choose to create?