Key developments in AI governance, regulation, accreditation, and enforcement, synthesized so you can stay informed without drowning in jargon.
Dense reports and frameworks from NIST, ISO, the EU AI Act, McKinsey, MIT, ECRI, EDUCAUSE, and the Harvard Business School translated into the decisions a board, CEO, or managing partner can use this week.
Original analysis on governance architecture, decision authority, workforce identity dynamics, and the human systems that determine whether AI investment produces value or unmanaged liability.
Not in the model. In the human architecture no one is designing.
Each edition examines the layer underneath AI adoption: who holds decision authority when AI is wrong, where the Human Authority Line is drawn for high-risk workflows, how organizations respond in the first ninety minutes of an AI-related incident, and what regulators, accreditors, and boards are actually measuring against. Concrete patterns. Named gaps. Specific moves you can make before the next board meeting.
Five questions that turn a complete board packet into real judgment: what carries the most weight, what is not surfacing, who can pause the system, what threshold triggers a review, and where speed should never be the priority.
The five categories every strong AI board packet needs: inventory, outputs, human intervention, incidents, and stress tests. Visibility is the floor. Judgment is what a complete packet still asks the board to bring.
Ford rehired more than 350 veteran engineers after its AI-driven inspection systems failed to deliver, then won its highest quality ranking in 16 years. What capability retention protects, and why it is a governance question, not just an operational one.
This month's through-line: accountability shifting upstream. A federal executive order steps back on AI oversight while states move the other way, and healthcare, higher education, and private equity all feel the same pressure from different angles.
Waiting for the AI landscape to settle before building governance only guarantees the same months of lead time later. Why the delay is a psychology problem, not a knowledge one, and the one question to ask every vendor this week.
Graduates are rejecting "just embrace AI." Forty-two percent already believe it will shrink their opportunities, and entry-level hiring in AI-exposed roles has already dropped. Why the booing was a signal, and how leaders keep junior talent when the first rung disappears.
This month's through-line: governance accountability. New data on physician AI adoption, state-by-state accreditation pressure on universities, and Colorado's AI Act reversal all point to the same shift, from whether to use AI to who answers when it fails.
The Blind Spot is written by Dr. Tiffany Masson, Psy.D., founder of Falkovia, forensic and clinical psychologist, and founding university president who built a medical school from a blank page under accreditation and regulatory scrutiny. The newsletter draws on what she sees inside live engagements with healthcare CEOs, university presidents, and venture capital and private equity firms governing AI in regulated institutions.
Every Sunday. One edition per week, structured as a single essay you can read in five to seven minutes.
Executive leaders in healthcare systems, higher education institutions, and venture capital and private equity firms carrying institutional, regulatory, fiduciary, or reputational risk. The audience is concentrated in the CEO, President, Chancellor, CMO, CIO, COO, General Counsel, and Managing Partner seats.
Most AI newsletters cover models, tools, vendors, and benchmarks. The Blind Spot covers the layer underneath: who holds decision authority when AI is wrong, where the Human Authority Line is drawn, how governance holds under regulatory and accreditation scrutiny. The frame is the human architecture, not the technology.
One of three content types: AI news distilled (regulatory developments, accreditation policy, enforcement actions), research made actionable (translating dense frameworks into executive decisions), or original thought leadership on governance architecture. Editions rarely mix types.
Yes. There is no paid tier. Every edition arrives in your inbox at the same cadence whether you subscribe today or have been reading since the first issue.
Yes. Every edition includes a one-click unsubscribe link at the bottom. There is no friction, no exit survey, and no follow-up email.
The Human Authority Line is the documented boundary where algorithmic recommendation ends and human judgment stays non-delegable. The Blind Spot returns to it often, because AI failures often trace back to authority that drifted to the system by default rather than by decision. Each edition shows where leaders are drawing that line, by workflow and by risk level, before a regulator or accreditor asks them to.
Whether the organization can show that its AI-assisted decisions are defensible: who approved each system, who can override it, what happens in the first ninety minutes of an AI-related incident, and what leadership would hand a regulator who asked to see the governance tomorrow. The Blind Spot frames these as items a board can put on its next agenda, not abstractions.
Yes. Editions track state AI legislation, CMS and accreditation developments, the EU AI Act, and the NIST and ISO frameworks, then translate what each one means for the specific seat the reader holds. The throughline across sectors stays the same: efficiency is the floor, defensible human governance is the ceiling.
A named person with real-time visibility into consumption across every deployed AI system, and the authority to act on what they see. Approval at the policy level does not govern spend unless someone owns the number day to day. The Blind Spot covers how leaders build that visibility and named authority into the architecture before an unexpected invoice forces the conversation.
A policy describes what should happen. Governance is whether the structure exists when someone applies pressure: who can pause a live system, who can reconstruct how a decision was made, who answers when a tool is wrong. The distance between what an organization can document and what it can defend is where its exposure lives.
A named person with pre-authorized power to stop a high-risk system immediately, not a committee. When everyone owns the risk, no one owns the decision, and the response window for a live AI failure is measured in minutes. The Blind Spot calls this kill-switch authority, and pairs it with the cultural guarantee that pausing a system for safety is not punished.
The failure point is rarely the technology. It is the human architecture that was not built before deployment: unnamed authority, rigid workflows, and a workforce responding to feeling replaced rather than extended. The tools tend to perform as designed while the organization around them stays structurally unprepared. The Blind Spot examines what trust architecture requires to close that gap.