AI-assisted hybrid VR training for demanding skills
Some skills are used only when the incident is already serious: specialist emergency response, and the equivalent niche skills in defence — hazardous-material and pollution response, harbour and littoral protection, combat-engineering tasks that cannot be drilled every week. The skill fades between events. Courses are fixed-date and costly. Every live exercise ties up equipment, crews and shift cover. The problem is not a shortage of slides. It is knowing who is ready before the scarce exercise day.
Akela has run this hybrid path with Finnish rescue services in Pirkanmaa, Lapland and Central Finland, first on oil-spill response, and built so the same pattern fits other rare specialist skills in civil protection and defence. Remote theory, a practical exercise on the unit’s own equipment, an after-action review, and a final exam. A certificate only after both the exercise and the exam. In one region the training was used in a real emergency: an incident you can call an emergency at −25 °C.
The authoring claim is the point for this audience. A specialist skill can be wrapped in a hybrid path and a browser scene from the unit’s own material in a day; the live exercise is then spent on the gaps, not on orientation. At ITEC this will be shown, not asserted: with permission, stand material — a brochure, an equipment photo, a procedure card, a short interview on what fails on the day — is turned overnight into a four-step path, one decision scenario, one 360 or photo-plate, and a sample gap brief. Labelled as a 24-hour authoring trial, not as that exhibitor’s course.
In a ten-plan pilot, completion was a poor proxy for competence: designed study time and time actually spent diverged, and calendar time to finish the self-paced theory stretched over many weeks around duty. The weakest criteria were not equipment handling. They were command, notification and prioritisation — the same gaps that appear in any rare specialist skill, civil or defence.
That is what the AI and the browser scene are for. The AI reads the planning task and returns a gap brief the trainee can act on and the instructor can confirm: reasons, feedback without points, a risk note, and three focus points for the exercise day. It does not certify, does not rate safety on the day, and does not see names or unit identifiers. Browser VR and 3D, built from the unit’s own photos and procedure, run on a laptop, tablet or headset, with no app-store lock-in and no field augmented reality. Neither replaces the instructor or the live exercise. They make the scarce day about the gaps, not about orientation.
Next step: hybrid course plus AI gap brief versus the standard course, scored on the exercise day, not on clicks. Seeking partners in emergency services and defence training for that comparison, and for local scenarios that are hard to reach, hard to control, or must be planned with limited resources.