Top 5 reasons AI tools tend to create predictable crisis scenarios
We're privy to some pretty illuminating research on the use of AI in training simulations.
F'rinstance, one top-line found that many executives imagine AI would be great at devising crisis scenarios and situations. And that's true, AI tools could generate crisis scenarios almost in seconds: They can trawl through historic patterns, connect causal events and produce a plausible narrative about how things might go wrong.
For crisis and communications teams under pressure to make more of their crisis prep efforts with less resources, that seems to present an attractive proposition.
But, as Matt Bevan finds in a recent episode of "If You're Listening"", AI has an innate weakness. Yes, AI is good at working with patterns. But way less potent when the challenge is to consider elements that are truly left-field in nature.
And that matters enormously in crisis simulation. For when we only train for what we expect, or what went before, we're open to ambush by maverick or renegade factors, which is what really tests the quality of our thinking and triaging capabilities. Great as it is, AI models remain - as Matt Bevan points out - eternally stuck in the past and, possibly, the predictable.
I reiterate, the value of crisis or risk simulations is not simply to rehearse for threats we - or AI - can readily imagine. It’s to expose the assumptions we didn't realise we were making, push leaders out of their conceptual comfort zones and test their prowess when events take a very surprising turn.
If AI draws predominantly on what has happened before, there's a clear likelihood its scenarios will be pertinent, plausible and reassuringly familiar. They might challenge an organisation's response plan, yet without ever challenging the thinking underpinning it.
Further, if we're all using AI tools trained on similar patterns, we may end up seeing variations of the same crises, leaving genuinely disruptive scenarios altogether unexplored.
Be clear; that’s not an argument against using AI. We use AI in our Drill simulations. Used well and in right proportion, it can accelerate scenario development and help teams explore the consequences of different decisions. Now back to those top 5 reasons AI tools may likely create predictable scenarios.
1. AI learns from existing material that may be outdated.
2. AI may not know the latest developments.
3. It reproduces familiar patterns rather than ensuring present-day accuracy.
4. It’s not a great fact-checker; outputs can contain outdated or incorrect details.
5. Without current, specific ‘human’ guidance, AI defaults to familiar examples.
You see, the real value of a crisis exercise isn't whether the scenario looks realistic (or how quickly your trainer devised it). It's whether the scenario actually prepares your people for when reality stops behaving as you expected and, thus, within your comfort zone.
AI can help realistically rehearse the foreseeable; and we can certainly do that in The Drill. The big challenge is ensuring we don't lose ability to realistically test for the unpredictable.
#CrisisManagement #ScenarioPlanning #AI #TheDrillcrisissimulator #simulationsoftware #IfYoureListening