(Possible) Rise of Artificial Intelligence for Resilience
All fields of science have evolved
since the rise of computers. Computers have helped researchers and
practitioners facilitate more accurate and
prompt answers to key resilience questions. Satellite imaging and remote
sensing have made it easier to track ecological changes (e.g. land cover) happening in an SES. Software programs have made
it faster and more accurate to forecast major disturbances (e.g. typhoons). Computer-facilitated
modelling tools have made it possible to project most-likely alternate states
of large SESs under certain assumptions (e.g.
climate scenarios). With the swift pace of modern learning curves, this
convergence of resilience and computer sciences is expected to increase over
the next few years.
In particular, the rapid development
of artificial intelligence or AI will highly influence the direction of
resilience science. It is not impossible that all these technologies currently
employed will soon be packaged into one Ultimate
Resilience Tool- a
centralized super computer with the capacity to analyze multi-layers of
datasets from databases and on-ground conditions.
This tool may be able to identify what
internal drivers to regulate, what disturbances to give extra preparation for,
and what to expect with nature’s ecosystem services in a future full of
Imagine this tool simultaneously analyzing
hundreds of internal and external determinants of the resilience of SESs, including major and minor ones! Imagine this tool
processing thousands of SESs data on land use patterns, demographic shifts,
climactic fluctuations, market dynamics, intergenerational interactions and many
others! Best of all, imagine this tool telling what most probable alternate
state the SESs is about to cross, when it
will cross it, and what happens when it finally crossed!
If this tool may indeed become a
reality, it has huge potentials to help local SESs become resilient against
impacts of global environmental changes. For example, this tool could have
immediately warned the hypothetical village that the number of their population
already poses a threat to their ground water. Or, that the intensity of an
incoming drought may be excessive for their existing SESs. It could save the
village from risks of leaving its year-round agriculturally productive state. It
could save the local SES from crossing an alternate state which offers fewer ecosystem services to support the people.
However, it is recognized that the
development of such tool may also threaten the people working with resilience
science. Currently, assessments of the resilience
of various SESs are conducted manually using integrated methodologies from
social and ecological sciences. Teams of researchers and practitioners go to
target SESs to collect onsite data. Surveys,
field samplings, and tons of secondary data collections are done for months.
But who needs to hire a whole team of
researchers and practitioners when a single tool can provide answers to the
same questions? Indeed, having an Ultimate
Resilience Tool may be a necessary evil for the whole academic field. It is
hereby argued that tapping the future of AIs for this kind of endeavor is
needed given the rate of global environmental changes. Local SESs are the most
vulnerable to impacts of these changes. In turn, this will greatly affect the
ecosystem services local people have long depended on. Thus, all options,
especially AI, should be harnessed to make these SESs resilient.