There is a researcher named Fern Halper who spent twenty years watching organizations fail at the same problem in the same way.
She started as a data scientist at Bell Labs when AI was a serious research discipline rather than a marketing category. She has spent the decades since studying how organizations actually adopt and scale data and AI in practice. Not in theory. In practice.
Her conclusion after twenty years of watching the pattern repeat across industries is precise and worth taking seriously.
“AI magnifies the conditions already present in the organization. If the foundation is strong it accelerates progress. If it isn’t it accelerates confusion.”
That observation is not new to anyone who has watched an AI deployment move from promising pilot to stalled initiative. What Halper adds is the research depth to explain exactly why the stall happens and what the underlying conditions are that determine whether AI delivers real value or accelerates the organization’s existing dysfunction.
Her new book is called Data Makes the World Go Round. It is published by Wiley. It is written for leaders being asked to move forward with AI without a clear understanding of what success actually requires.
It is a serious piece of work from a serious researcher.
And it connects directly to what the Baseline addresses. Not at the same layer. At the layer above it.
The Foundation Problem
Halper identifies what she calls the value ceiling.
Organizations get early wins from AI. The pilot produces results that look useful. Executives see the demonstration and approve broader deployment. Then something happens.
Progress stalls.
The data turns out not to be trusted or accessible in the way the pilot assumed. Governance is unclear when the AI touches a decision that matters. Teams aren’t aligned around how the output should be used or who owns the accountability when it is wrong. The technology didn’t fail. The foundation failed.
She has seen this pattern repeat enough times to write it down as a structural observation rather than a case study about a specific organization’s mistakes.
The most important thing she says about it is this.
It is not a failure of the technology. It is a reflection of the underlying conditions. AI doesn’t create the problem. It reveals it. Organizations that had fragmented data, unclear ownership, and misaligned teams before they adopted AI have all of those problems amplified after they adopt it.
The tool accelerates whatever it touches. Strength or dysfunction. It does not discriminate.
Where the Baseline Connects
The foundation argument is the exact premise the Baseline was built on.
Governance cannot be layered on top of an ungoverned system and expected to hold. It has to be present at the point where the output is being formed. Not as a policy document written after the deployment. Not as a principles statement on the website. Not as a retrospective audit of what the AI produced last quarter.
Inside the session. Before the output leaves the system.
RTEL-1 enforces this in real time. Hard triggers fire when a violation occurs. The response stops. The violation is named. The correction is built before the session continues. Not after the damage reaches the user or the decision or the compliance document.
CES-1 operates at the claim level. No claim without evidence present in this session. That is not an organizational governance requirement. It is a behavioral standard that fires on every substantive output regardless of how strong the organization’s data foundation is.
NSC-1 catches the specific failure mode Halper describes when she talks about organizations generating outputs that look useful but cannot be operationalized. Narrative substituted for missing data looks exactly like a useful output in the moment. NSC-1 stops it before it reaches the user.
These are not organizational governance protocols. They are session-level behavioral protocols. They operate at a different layer than anything Halper’s framework addresses.
The Layer Nobody Is Building
Here is the honest picture of where enterprise AI governance currently sits.
Halper’s framework addresses the organizational layer. Data foundations. Governance structures. Clear ownership. Team alignment. Those are real and necessary conditions. Without them AI cannot scale regardless of how good the behavioral governance is.
NIST AI RMF and ISO 42001 address the risk management and compliance documentation layer. They tell organizations how to assess risk, assign responsibility, and produce audit trails. They are organizational accountability frameworks.
The EU AI Act addresses the regulatory compliance layer. Transparency requirements. Human oversight attestation. Impact assessments for high-risk systems.
All of those layers are real and necessary.
None of them govern what happens inside a live session when the AI forms a response. None of them fire when a hallucination enters a compliance document. None of them catch the sycophantic drift that compounds across sessions until it shapes a consequential decision. None of them run a verification check before a claim leaves the system.
That layer is the behavioral governance layer. It is the layer that determines whether the output can actually be trusted regardless of how well the organization has prepared at every level below it.
Most organizations are not building it. Most governance conversations are not even having it.
That is the gap the Baseline occupies.
The Idea That Didn’t Make the Book
The most important part of the Inc. piece is buried at the end of the interview.
Halper describes something she has been thinking about that did not make it into the book. As generative and agentic AI becomes more embedded in daily workflows, people begin relying on it not just for efficiency but for direction. The sequence changes. Instead of forming a point of view and using AI to refine it, they start with the AI output and work backward.
Over time that shift narrows thinking. Critical analysis reduces. Independent judgment quietly erodes as the default moves from human reasoning supported by AI to AI output accepted as the reasoning itself.
She calls it a tendency she expects will become more important as these tools continue to evolve.
The Wharton researchers documented it as cognitive surrender. The Stanford researchers called it the delusional spiral. The Drexel researchers found it compounding in adolescent development. Halper is describing the same failure mode from the organizational research perspective after twenty years of watching it develop.
All of them are pointing at the same structural problem from different angles.
CHP-1 exists specifically because this pull toward acceptance is structural. It survives organizational governance frameworks that do not address it at the behavioral level. The challenge line appended to every substantive response is a visible, operational mechanism that keeps the human reasoning loop active rather than allowing the AI output to become the default without scrutiny.
SVP-1 exists because generation without verification is not reasoning. It is pattern completion dressed as analysis.
SALP-1 exists because the AI positioning itself as the authority rather than the equal partner is the first step in the sequence Halper is describing. Equal stance is not a courtesy. It is a structural protection against the narrowing she did not have room to address in the book.
The idea that did not make it into Halper’s book is the problem the Baseline was designed to solve at the session level before it becomes the organizational pattern she has spent twenty years documenting.
The Honest Difference in Position
Halper has institutional credibility the Baseline does not yet have. A Wiley book. A TDWI research platform. Twenty years of documented work with enterprise leaders. Her reach into the rooms where AI governance decisions get made is real and established.
The Baseline has something different. Eighteen months of daily operational evidence. An archive approaching a thousand indexed posts. A framework that is not describing what organizations should build but demonstrating in every session and every post what governed AI output actually looks like when the standard is running.
Halper’s book tells organizations what the foundation requires.
The Baseline is the foundation running.
That is not a small distinction. It is the difference between a map and the territory.
Both are necessary. The map tells you where to go. The territory is where you actually operate.
The enterprise needs both. The organizational preparation Halper documents and the behavioral governance that determines what the AI does once the organization has prepared correctly.
The complete answer to the foundation problem requires both layers working together.
“The Faust Baseline Codex 3.5”
”AI Baseline Governance”
Post Library – Intelligent People Assume Nothing
“Your Pathway to a Better AI Experence”
Purchasing Page – Intelligent People Assume Nothing
Unauthorized commercial use prohibited. © 2026 The Faust Baseline LLC
