AI Misadventures And The Adversarial Economy
cxotalk.com, Wednesday, August 6th, 2025
Why AI fails in the real world: the adversarial economy-opposition, change, wrong objectives-with cross-industry cases and early warning signals. Learn from two brilliant speakers.
AI's biggest failures aren't software bugs. They're market outcomes.
Episode 890 examines why AI stumbles when it encounters an adversarial economy, where incentives collide, conditions shift faster than models can adapt, and determined actors, including malefactors, learn to circumvent controls.
We organize the discussion around three forces:
- Opposition: People exploit ranking, identity, pricing, and support flows
- Change: Policy, behavior, and data mix move and models lag
- Wrong objectives: Metrics and design choices reward the wrong outcomes
Our guests include Anthony Scriffignano, a seasoned leader in global data integrity and risk, and Steve Daffron of Motive Partners, whose focus spans complex, technology-driven enterprises. Together, we explore how incidents originate, why they spread across vendors and borders, and why dashboards can remain green even as harm increases.