The brain science of sexual harassment prevention training pt2

In light of recent high profile harassment allegations, Todd Maddox thinks training can be the key to prevention, if you get it right.

First-person interactive branching procedures

An approach that is gaining popularity involves ‘branching’. As with so many things, one must be careful to evaluate the specific offering since the same term can be used for different training approaches.

In one type of branching scenario, the learner is in a first-person CBT or VR setting and is involved in some situation as it unfolds. At some point, the experience will pause and the learner will be asked what he or she would do next.

The learner’s response will determine how the scenario plays out. This is definitely a step in the right direction. The learner has a first-person perspective, and with VR this will be immersive. The learner is an integral part of the scenario and is experiencing behaviour first hand.

The main weakness of this approach is that the real-time interaction stops, and instead of training behaviour, the learner’s cognitive skill repertoire (i.e., their ‘what’ knowledge) is queried. Although asked what they would do, one is very likely to respond with what they know that they should do.

From a brain science of learning perspective, the key to behaviour change is real-time interaction.

Although the detailed neurochemistry is beyond the scope of this article, suffice it to say that behavioural skills are learned gradually and incrementally via dopamine-mediated, error-correction learning in the basal ganglia of the brain.

The key, and what all developers must keep in mind, is that behaviour change is only effective when the behavioural skills learning system in the brain is effectively recruited.

When a behaviour is generated that is rewarded by a smile or a head nod, the neural connections that led to that behaviour are strengthened and that behaviour is more likely to happen again. When a behaviour is generated that is punished by an angry face or a shaking head the neural connections that led to that behaviour are weakened and that behaviour is less likely to happen again.

Learning in the behavioural skills system in the brain is active, it involves learning by doing, and physical repetitions. The approach to branching that breaks this real-time feedback loop reduces the chance of behaviour change.

A more promising approach, and one that would be ideal from a brain science of learning perspective, would be to ‘branch’ the scenario based on the learner’s behaviour, but in real-time. 

This could be achieved through an artificial intelligence (AI) approach in which the learner’s behaviour elicits an AI-driven response that rewards an appropriate behaviour on the learner’s part and punishes an inappropriate behaviour on the learner’s part.

Alternatively, live actors could be trained to drive the behaviour of avatars in simulated CBT or VR worlds. As technology progresses, the avatars will get more realistic, which will only enhance the effectiveness of the tools.

The key, and what all developers must keep in mind is that behaviour change is only effective when the behavioural skills learning system in the brain is effectively recruited. The critical factor in recruiting the behavioural skills learning system is real-time interaction.


The learner’s behaviour must be followed, literally within a few hundred milliseconds, by corrective feedback. When this necessary feature is present, behavior change will occur.

The promise of true interactivity

A harassment prevention training tool that takes as its central tenet real-time interaction is the key to success. It also holds significant promise in a number of other ways. First, one can include a broad set of scenarios during training.

Broad-based training enhances transfer and generalisation to related – but untrained – situations, and will generalise beyond harassment to other aspects of people skills. Although sexual harassment is clearly an issue, so is harassment based on sexual orientation, race, ethnicity, disability, and age.

Second, one can focus on behaviour change in the potential harasser to decrease the likelihood of inappropriate behaviour, but one can also provide tools to the potential victim to help them better de-fuse the situation. Finally, bystanders can be trained on the appropriate methods for diffusing potentially harassing situations.

Although bystanders may not be present during the most egregious acts, there is good evidence to suggest that harassers often start with milder harassment behaviour. More responsive bystanders could reduce the likelihood that potential harassers go to the next level.


Interactive VR platforms are ideal for harassment prevention, and soft skills behaviour training because the learner is immersed in the situation, they have the feeling of presence, and their behaviour are being trained in real-time.

In addition, if the platform is constructed appropriately, the learner can be placed in a broad range of situations and scenarios that include small or large numbers of bystanders, a broad or narrow array of ethnicities and genders, etc. Broad-based training enhances generalisation and transfer and trains the individual to be prepared for any situation.


About the author

W. Todd Maddox, Ph.D. is a contributing analyst at Amalgam Insights, Inc.


Learn More →

Leave a Reply

Your email address will not be published. Required fields are marked *