Governance told you who decides.
Nobody engineered what deciding actually means.
Financial Services · Healthcare · Boards · Corporate Governance · Institutional Accountability
Issue #003 moves to Layer 3 — Intent. Where strategy is translated into operational direction, and where governance frameworks rarely look. Subscribe to receive it when it publishes.
The cigarette had a good run before anyone called it what it was. Medicine, they said. Good for nerves. Good for digestion. Handed to soldiers so they could focus. For a long time, nobody was lying exactly — they just weren't asking the right questions.
Then came the autopsies.
AI is on the same clock. Marvel → ubiquity → hidden costs → reckoning. We are somewhere between stage two and three. And we are still inhaling.
Cigarettes once fixed digestion, calmed nerves, cleared airways — or so the marketing went. Now it's AI solving climate, curing cancer, and eliminating inefficiency. The pitch doesn't change much. Boundless upside. Cost deferred. Someone else's problem, later.
At some point, the cigarette stopped being a product and became part of the furniture. AI got there faster. It's already in hiring decisions, loan approvals, medical triage, and content feeds. Invisible infrastructure. And opting out stopped being possible some time ago.
The hardest thing to explain about cigarettes wasn't the lung cancer. It was the non-smoker in the next room. Algorithmic second-hand smoke is already accumulating — distorted information, quietly displaced workers, attention systems rewired without consent. You don't have to use AI to be changed by it.
One report in 1964 changed what the cigarette meant. Not the science — the science had been building for years. Just the official admission. The EU AI Act is trying to be that moment. The warning labels are being drafted. Meanwhile, deployment doesn't wait.
The industry's answer to harm was the filtered cigarette. Same nicotine, less tar, cleaner story. AI's version is the small language model — tighter training data, closed systems, the toxic layer stripped away. The generation currently growing up inside these systems will settle that question. Not us.
Initial utility was never proof of long-term safety. The cigarette proved that. A generation got sick before the institutions admitted it. The question now isn't whether AI has second-hand effects. It's how long we wait before we start counting them.
The arc: miracle → infrastructure → harm → reckoning → refinement.
We are entering the harm phase. The reckoning hasn't started.
The tobacco companies were also building factories in the 1950s. Investment doesn't pause harm. It scales it.
No theory. No governance abstractions. Real failures, real institutions, structural fixes you can take into your next board meeting.
Subscribe on Substack — Free →// Academic Reach
Six papers published on SSRN. Eleven eJournals. Covering institutional decision integrity, fiduciary risk in automated systems, corporate governance architecture, and decision drift in banking and healthcare. Research developed alongside faculty from Johns Hopkins, Harvard, Fordham, the University of Zurich, the University of Toronto, the University of South Carolina, and Purdue.
I work at the intersection of institutional decision architecture, corporate governance, and fiduciary accountability — in banking and healthcare, where the consequences of a broken decision chain are measured not in inconvenience but in regulatory breaches, capital erosion, and loss of trust.
For over two decades I have watched the same pattern repeat: institutions modernise faster than the governance architecture connecting board intent to execution. The gap between what a board mandates and what the institution delivers has always existed. Automation made it invisible at scale — and consequential at speed.
Decision Engineering™ is the discipline I developed to close that gap — not through more governance, but by engineering the decision chain itself.
I take a small number of engagements each year, working directly with banking and healthcare leaders on decision architecture, AI governance, and institutional accountability.
If you are reading this newsletter and recognise the pattern in your own institution — that is usually where the conversation begins. Reply to any issue, or reach out via Substack with the subject line: Decision Integrity.
Reach out via Substack →