In partnership with
Dear {{ first_name | reader }},
This week's article leans on a study from IBM's Institute for Business Value, and a disclosure is worth making before you read another word: I spent ten years inside IBM earlier in my career as Public Relations Manager.
I still read their research through that lens. Not a conflict of interest but a bias worth naming rather than leaving you to guess at it.
That history aside, the numbers stopped me because they describe a crisis type most of us haven't built a plan for. Not the kind where AI generates a bad citation or a convincing deepfake. The kind where an organisation discovers, mid-disruption, that it never mapped what it was standing on in the first place.
So this week: three ways AI can sit at the origin of a crisis, why only one of them looks like the crises we're used to planning for, and what a global study of 1,000 executives says about how exposed most organisations actually are.
Enjoy!
WAG THE DOG NEWSLETTER | ISSUE WEEK 27, 2026
KEY TAKEAWAYS
AI can be the origin of a crisis in three distinct ways — by causing it directly, by complicating a crisis that started elsewhere, or by exposing an organisation's own structural dependency on AI systems it never fully mapped.
91% of surveyed executives don't fully understand their organisation's AI dependencies across vendors, models, and infrastructure, according to IBM's Institute for Business Value.
Structural dependency exposure is the least planned-for origin type because it has no bad actor, no discrete failure, and no clear trigger date to point to.
Six AI-related disruptions hit the average surveyed organisation over the past two years, yet 81% say a seven-day vendor outage would still cause severe or critical disruption.
Only 7% of organisations operate with advanced AI control capabilities, and that group protects 55% more operating profit from AI-driven disruptions than the rest.