Thinking out loud on governance, structure, and what it takes to stay in control. Where AI is used without governance, organizations don't just take on risk. They lose control of what they are becoming.
Christopher Trocola of AICT has published research that should change how organizations think about their AI workforce. It documents large-scale technology job losses with AI cited as an explicit driver, rapid adoption of AI for resource allocation across US law enforcement agencies, and a legal reality most organizations have not absorbed: AI systems used in hiring, lending, housing, and benefits determination are already subject to existing statutes, including the Equal Credit Opportunity Act, the Fair Housing Act, and Title VII, and courts have begun issuing adverse rulings against organizations that cannot document how their AI systems reach decisions.
The report establishes the environment. It does not answer the question every organization will soon need to answer.
When a regulator, a plaintiff's counsel, or a board asks who in your organization is accountable for the decision an AI system influenced, what do you say?
Most organizations cannot answer that question cleanly today. The training programs being purchased to address this do not produce an answer either.
Your AI vendors are counterparties, not just products. Their financial durability is a risk most governance frameworks don't assign to anyone.
Continue on Substack →A short, practical way to check whether your organization's AI governance would survive contact with a regulator, an auditor, or an incident, before you find out the hard way.
Continue on Substack →The fastest-growing governance risk isn't the AI your team adopted. It's AI-written code no human read, shipping inside what you buy, into your regulated environment.
Continue on Substack →The vendor holds the switch that turns your AI system off. If you have never tested what happens when they flip it, you do not have a contingency plan. You have a hope.
Continue on Substack →Running a decision through an AI does not make it the AI's decision. Two regulators have proven it, with a bill attached. Decision-laundering has no legal standing.
Continue on Substack →Firms say they're waiting for the AI rule that already arrived. Regulators withdrew the wait, not the obligation, and what counts as compliant was written down while everyone was still stalling.
Continue on Substack →A large share of what's sold as AI is performance, backstopped by people. At Presto, "human in the loop" wasn't a safeguard, it was the labor that hid the gap. The buyer inherits it.
Continue on Substack →Every layer of AI security limits what the agent can do or watches what it did. None governs whether the action was right. That gap is not a missing control. It is a missing owner.
Continue on Substack →Accountability, and the control structures organizations need before AI starts making decisions they didn't authorize. Every edition raises a question worth pressure-testing. If it surfaces something you're navigating inside your own organization, or you see it differently, write back at media@fellowshipintelligence.com.
