How AI could alter EHS training forever

Intro

In today’s workplaces, people need to constantly renew their skills and knowledge. Equipment is updated, processes evolve, and legal, social and economic requirements change priorities. A “one-off” education or training course can’t meet these needs. Elearning has filled the gap successfully in some places, with organisations able to roll out short courses quickly, so that all employees can be presented with information about a new hazard or new requirements.

But conventional elearning is struggling to keep up with the pace of change, the changing nature of the workplace and the workforce. Rather than applying a known set of rules to a known set of conditions as in the past, workers need to understand principles and apply them in new ways to changing environments.

Forward-thinking organisations with significant training budgets have been experimenting with virtual reality (VR) for EHS training and awareness. For example, the Port of London Authority used VR to enhance induction training. Learners made decisions during the experience which changed what happened next, leaving them with a better understanding of the impact of their choices. However, VR is not yet widespread, partly because of the cost of producing good learning materials and partly because of the limitations of the experience.

The convergence of artificial intelligence (AI) with VR can overcome some of these limitations, making better VR more cost-effective to produce, and more realistic to use.

Creating great VR

With regular VR, every route through a scenario must be considered, defined and then either filmed using 360° cameras with real people and places, or animated by graphic designers. The VR can only let you experience the scenes that have been filmed or animated. As a user, it is frustrating to be stuck in a VR scenario which won’t let you continue until you make the selection the designer expected. One benefit of VR should be to let people make the wrong decision to find out (safely) what would happen, or to see if there is an alternative right solution. For decades, airline pilots have been able to do this using flight simulators, but such solutions are disproportionately expensive for most industries.

With machine learning, VR will be able to create new routes through the experience when learners respond in ways that the designers hadn’t foreseen. Rather than needing human animators or actors to create new scenarios, AI can use knowledge about how people move, how equipment behaves and what objects look like from any angle to create new imagery as a learner decides how to move around their virtual environment. In the past VR has been based on gaming technology – you only have to watch a ’Mario’ style game to see how the rules of physics are ignored. Many VR fire-fighting and evacuation scenarios show fires growing too slowly, smoke behaving too uniformly, and fire extinguishers acting too effectively. Another form of AI, computer vision (CV), can learn from watching real behaviours of people and workplaces. If this is applied to VR scenarios, people and hazards could be modelled more realistically.

Controlling great VR

Some early attempts at using VR for EHS training have resorted to using pull-down text menus overlaid on images to give users choices. Learners have to manipulate hand-held controllers to select options in the menu. The appearance of text, and the need to remember which button points and which selects spoils the immersive experience.

Advances in CV are already allowing learners to get rid of the controllers and use their hands in more natural ways to control the experience, something which is likely to accelerate in the next year or two. 

Natural language processing and speech recognition is an area of AI that has been promised for decades, but only now is the processing power becoming available that will allow learners to direct the VR experience using voice. “Chatbot” style AI will give the VR characters the ability to talk back so that learners can have realistic conversations with characters in the VR. This will make VR a great tool for teaching the non-technical skills that are so important in EHS – such as how to have positive safety conversations, how to explain safety rules, and how to have problem-solving conversations. 

Conclusion

Although VR has been shown to be a useful tool for EHS training, the cost of producing it by conventional methods and the limitations of using it keep it out of the hands off many organisations. Developments in AI such as computer vision will allow great VR to be created more cost effectively, and to be used with a greater impact on EHS training. AI applied to VR could alter EHS training, forever. To learn more about how Protex AI is using vision to help EHS safety managers develop a proactive safety culture, chat to one of our product experts here 👈🏼

‍