Generative artificial intelligence is transforming crisis management. It helps situational
awareness and decision-making by processing real-time data quickly and making more
accurate predictions. It speeds up the planning and response process, refines evacuation
plans, optimises resource allocation, creates crisis scenarios if needed, provides
estimates
for damage management and overall consequence management.
However, it not only analyses (provides computational assistance) but also provides
room
for improvement. It also contributes to the development of capacities and capabilities
by
providing a venue, a role-based virtual space for crisis training. The various AI-driven
crisis
simulations provide a low-risk space for crisis training where crisis teams can exercise
in a
(near) real-life situation without the risk of suffering real damage. It also provides
a way to
gain experience, learn from mistakes, increase self-confidence and reduce stress reactions
during crises.
AI can therefore be used for crisis modelling, forecasting, risk reduction, human
capacity
building, measurement, scaling, real-time threat monitoring in combination with other
tools (e.g. geospatial tools, unified emergency hotlines or sensors). My research
question is
that, despite its many already tangible benefits, what artificial intelligence does
not solve,
what will continue to be and is expected to be for a long time to come to be solved
by
human actors, crisis and other experts in crisis management. What do machines not
have
the answers to in the age of learning algorithms for crises?
My method is literature analysis and case study. I complement the overall industry
and
academic work with an analysis of good practice, and try to highlight the discrepancies,
which unsurprisingly are most apparent in the area of psychological responses to crises
and so-called soft skills. In crises, emotional intelligence, intuition and spontaneity,
creativity and inspiration, empathy and caring are all valued... And the key to crisis
management is usually what cannot be predicted.
