Artificial Intelligence is coming! Yet beneath the buzz lies a simple idea:

Spot patterns to make suggestions.

Here at GoodPractice the IT boffins think this idea can be applied to our leadership and management content.

Imagine a site that could learn who you are and invisibly adapt to your shifting interests. It would be a site that was just plain interesting. For each individual the content would be equally pertinent, but only by being subtly different.

Content for individuals vs organisations

GoodPractice has always strived to present truly useful content. To help us we use technology that allows us to craft bespoke sites for each individual client organisation. This works certainly, but it implies a compromise: it is difficult to craft a content structure that will satisfy individuals when the target is their organisation.

The naive solution would be to try, somehow, to target the individual user. In the past this has been done through personalisation settings. You may know the kind of thing: big lists of tick boxes or draggy-droppy content panels. This approach does not work because it is too much like hard work and also people change, their roles change and so their priorities change.

The solution is not burden the user with keeping their site settings up to date; instead, the site should be continuously adapting to you.

To achieve this utopian ideal we must first stop treating the user and their organisation as distinct things. Instead, we must view a user as a faceted being, a conglomeration of different aspects: we should understand the user as an individual and a member of an organisation, a manager and a colleague, a leader and a follower … all mixed together in one cuddly package.

And this is where the ability to spot subtle patterns comes in useful.

Ripples in the datasphere

ripples

Wherever you roam in the datasphere you are making ripples. The pattern of those ripples depends on your individual personality, your role within your organisation and also the shape of your organisation, its goals, culture and, indeed, anything else that contributes to the choices you make. Although the effect is subtle, the waves of an ocean are changed when you dip your toe in at the water's edge. The problem is that human minds are not suited to seeing these noisy, interleaving patterns - for that you need a machine.

So AI, when you boil it down, is not really about evil robot overlords clanking through the workplace; it's just the same old story of computers doing stuff that humans are not very good at, like spreadsheets or spelling or noticing deep patterns - only more so and much faster.

Spotting patterns

cheetah spots

Once the patterns have been spotted then it is relatively easy to categorise a person as being similar to other people and so start making intelligent suggestions. It's like that old TV show, Mr & Mrs, where a married couple answer questions about each other’s preferences for milk-in-tea or soft furnishings. Once you know somebody well enough you can take an intelligent punt on what they might like - even for novel categories.

Another application of Machine Learning is helping to make the learning stick in the minds of our users. Experimental evidence indicates that the act of retrieving new information makes it more likely to be remembered. [1]

Imagine you encounter the phrase:

‘information is stored in neuro-electrical patterns’

A few minutes later you are asked to fill in the blank:

‘information is stored in neuro-elec***** patterns’

Although this is trivially easy, the act of recall reinforces the neuro-electrical patterns in your brain. It’s not a test, it is not meant to be hard, it is about forcing the user to do just enough work to fire up the specific neural pattern and so both strengthen it and mark it as suitable for future recall - it’s brain-whispering.

Lighting up your neurons

Synapses

With this in mind, another rich topic of research here at GP labs is the dynamic generation of neural-reinforcement questions for each of the many items of leadership and management content in our library. The goal would be for a machine to intelligently parse the piece of content you are reading and then, using simple recall questions, cause the neurons to light up and so reinforce your recall. This is the efficient read-recall loop that shortcuts any number of boring and ultimately inefficient online multiple choice questions and assessments.

Artificial Intelligence, Machine Learning, Evolutionary Algorithms, all lovely words to make a geek’s heart sing, but really just tools to help our users learn.

It's all about efficiency using AI tech to optimise the presentation of our content, tailoring it to who you are both as a person and as a component of your organisation. This is exciting for sure, but making the content pertinent is only useful if you can recall what you learn. AI can help here as well. By making use of the latest neural research, machines can intelligently reinforce your learning as you go. Now learning can be interesting and useful.

Machines that augment humans - that's the real story of the AI revolution.


[1] Peter C. Brown, Henry L. Roediger and Mark A. McDaniel, Make it Stick: The Science of Successful Learning (Harvard University Press, 2014).

  • Robots image credit: Flickr (accessed 30 June)
  • Green seats image credit: Flickr (accessed 30 June)
  • Ripples image credit: Flickr (accessed 30 June)
  • Cheetah spots image credit: Flickr (accessed 30 June)
  • Synapses image credit: Flickr (accessed 30 June)