There has been extensive talk recently about systemic crises and polycrises and I have written a few posts on the topic (mostly in Italian). Separately I have also taken an interest in AI trying to get my head around the concept. Among the first to register on Chat GPT I never took the time to actually sit down and figure out what it is all about. What intrigues me is how we can apply AI to crisis management (specifically crisis preparedness).
So yesterday I decided to combine my interests and
after skimming through the recent World Economic Forum Global Risks Report 2023 released a few days ago I gave it a go. I must say upfront that I approached this exercise with great skepticism.
After a few warm-up questions on systemic crises and
polycrises I jumped right into it.
My first real question was “Looking ahead at a 10-year horizon scan, which systemic crisis scenarios do you anticipate?”: My intention was clear: compare the WEF Global Risks Report 2023 with Chat GPT’s analysis. The video will satisfy your curiosity. Suffice to say the bot suggested 8 systemic crises that it “thinks” can materialize making however clear that “It is important to note that the scenarios mentioned above are not exhaustive and that new systemic crises may emerge in the next 10 years that are not currently anticipated. Therefore, it is essential to continue to monitor global trends and to be prepared to respond to potential systemic crises as they arise.” A good lesson for all.
As the WEF closed-end questions were developed by the researchers we cannot draw an exact match between the “experts’” horizon scan based on known knowns and the AI’s. However, in the WEF’s report, 6 out of 10 areas of concern revolved around the environment whereas Chat GPT had crises spread across categories. In conclusion no significant differences here but some interesting nuances.
I then asked Chat GPT “Can you imagine an unconventional systemic crisis scenario likely to occur over the next 10 years?”. We could consider these falling between the “known unknowns” and the “unknown unknowns”. The reply here was in my view extremely interesting. All 5 scenarios presented were related to technology. Four out of five struck me. AI was (ironically) a given. But how about space, (the return of) genetic engineering, quantum computing, and virtual worlds?
The experiment has proven a point. As AI progresses and we learn to prompt the systems we may be able to quickly generate unconventional systemic crisis scenarios we can then examine, study and prepare for. What would have required considerable time and probably discussion with other professionals and experts was generated by Chat GPT in a matter of seconds. Everything must be taken with a pinch of salt but… my skepticism is over.
The problem of course is that we still need to completely rethink how we manage 21st-century systemic crises. But Chat GPT had a few ideas about that too!