E-learning and the Ethics of Robot Bullying

Reading Time: 4 minutes


Trilateral Research |

Date: 21 April 2022

When considering Ethical Artificial Intelligence we want to hold machinery to account and to protect humans from the potential harmful impact of powerful machines.  How transparent are the decisions that machines make?  Can we identify the bias that comes from selective availability of data?  Or from existing bias in the situations they are trained from?

Anyone who has watched the incredible Boston Dynamics robots might start to see this direction of ethical fit from a different view.  To what extent do machine processes need to be protected from the unethical behaviour of humans?  At what point does testing the capability of robots become, more simply, robot bullying?

If you lived in a world where self-driving cars were considered faultless, how carefully would you look before crossing the road?  When driving, would you give way to a self-driving car, which you know will slam on its brakes if you cut in front of it?

When designing training programmes, this issue of humans not respecting automated processes arises in e-learning.

Most of us have had to complete e-learning for compliance, perhaps cybersecurity or anti- money laundering training.  And “had to” is the tell-tale phrase – we have yet to meet anyone who has found compliance training enjoyable, and this negative mindset affects how we interact with it.  We want to get through it as quickly as possible, which means that any attempt to make the training more interactive, for instance by incorporating video narratives, can simply start to be viewed by the learner as an obstacle to progress.  How to get through the training in the quickest way becomes the approach of the trainee.

That mindset is the one that we most want to avoid, when designing training courses for behaviour change.  To some extent, having a human deliver the training is a large part of achieving this goal.  Social norms (hopefully) force us to interact politely with a human and, at its simplest, to pay attention.  Whereas when faced with something electronic, we are more tempted to become that person who steps in front of driverless cars or tries to trip up robots.

Interactive picture with clickable content?  Just click everywhere on the screen randomly until the content shows.  Narrative of several screens with a test at the end?  Just click through them all and try the test first, then if we pass, we don’t have to sit through the whole thing.  This kind of interaction is great for ticking boxes for compliance, but it won’t drive the behaviour change that is the holy grail of training programmes.  Mechanical, repetitive processes encourage us to short-circuit the reward process – anyone who has had to repeat a section of compliance e-learning and sit through the “fun” video a second time will identify with this motivation.

The answer lies in non-linear processes, and the most effective way to achieve that in both the learning process and assessment is through game play.

E-learning that takes the trainee on a journey of their own making gives them identity as a learner.  This is key to audience engagement – when any activity gives us identity is when we truly rise to that role.  It’s why everyone gets so nervous when they first perform public speaking, because our identity is being tested in front of our peers.  It’s why the best first-person computer games are so absorbing, because we start to identify with the character we are playing, so that the obstacles they face on screen are not something on a computer but something that is happening to us.

What does that look like when applied to a training course?  Game play is characterised by multiple pathways that create a narrative.  The trainee can explore a range of options that create the impression of a virtual world, incentivised by discovering or collecting rewards of some kind.  This will look different for every course, but to see it in action have a go at the excellent Lifesaver, which combines narrative with a multiple-path first person experience leading to serious virtual consequences.

One of the appeals of e-learning is the great cost saving over in person training, when delivered at scale.  Those who are prepared to reinvest some of that cost saving have another option open to them – human assessment.  This allows e-learning courses to incorporate free text answers, which greatly increases the complexity of the tasks that trainees can undertake.  This hybrid model gives back what we are prepared to put in, allowing just a little of the e-learning cost saving to provide a much richer experience overall.

If we were playing a game of football with that Boston Dynamics robot, would we be more respectful of it as we are having fun?  If we could speak to it with natural language and it could understand us, and we knew a human assessor might be watching, would our interactions with it become more sophisticated and constructive?

We have looked at two ways to draw learners into the e-learning process, through game play and hybrid assessment.  E-learning, like robots, is here to stay, and by changing the learner’s perspective from a course that must be completed to one that is fun and absorbing then we, as educators, can turn it into a tool that drives behaviour change as well as the next in-person courses.

Some of the best praise that any educator can receive for their course is for attendees to recommend it to their colleagues.  E-learning should be no different, and through game play and hybrid assessment we can get one step closer to that real success.  We build on that path to success with every course that we deliver, starting with our vision for success and working back from that to create the most engaging path to that success.

Game Play as simulation will comprise part of our HEROES project where border staff and transport workers are guided through realistic scenarios to identify victims of human trafficking.  By placing our trainees in situations that closely resemble real life, while immersing them in a realistic narrative with real-seeming consequences, we maximise the impact of behaviour change when they return to work to use these new skills.

Speak to us at Trilateral Research to find out how we can use these same techniques to achieve lasting behaviour change for you.

For more information, please contact our team.

Related posts

Let's discuss your career