Beyond the Forbidden Zone: Where AI tells us things we don’t want to know
What happens when an AI-powered curriculum management systems tell us everything?
The Forbidden Zone from The Planet of the Apes (1968).
In the original 1968 version of the film The Planet of the Apes, Taylor, the protagonist space traveler from Earth, is tried and convicted by the Ape Tribunal for the crime of being human and is sentenced to the custody of Dr. Zaius, the ape Minister of Science and Chief Defender of the Faith.
In a private audience, Dr. Zaius pronounces to Taylor that he is destined to be “emasculated” and then subjugated to experimental surgery on his brain to remove his ability to speak – “…a kind of living death.”
I offer an edited version of their exchange below, but instead of the characters of Dr. Zaius and Taylor, let’s pretend that it is a dialog between humans working in higher education and an embodied, well-trained and ambitious AI entity. In an unusual gesture of collegiality, Dr. Zaius addresses Taylor as a peer to find out where he really came from.
AI: “So, you don’t believe the prosecutors’ charge that I’m a monster created by Dr. Zira?”
Humans who work in higher education: “Certainly not,” admits Dr. Zaius, “You’re a mutant! I admit that where there is one mutant, there is another, and another, and another – a whole nest of them! You are a menace – a walking pestilence! I do know who you are, Taylor. I give you just six hours to make a full confession. After that, I shall use surgery to obtain one.”
AI: “All right, you can cut pieces out of me – you’ve got the power…”
Humans: “Guards! Return this creature to his cage!”
AI: “…but you do it out of fear! Remember that! Remember that!! Because you’re afraid of me!! What are you afraid of, doctor?!!”
This is a half-satirical, half-serious portrayal of the tension-filled interplay between the growing capabilities of AI systems and those of us in a knowledge industry where we feel we must either fight the damn beast or invite it to sit down in our kitchen. The collision of worlds is already here, but we are naïve to believe that the effects are limited to pivoting to AI-proof alternative assessment strategies.
If you know the plot of the film, the underlying context of the dialogue punctuates how this exchange has meaning beyond the words themselves. Taylor’s astronautical journey into deep space leads him to unknowingly return to Earth in the distant future. But it is not until the end that he discovers that ape society is what emerged after humans blew up the Earth with nuclear weapons. Until then, he thinks he’s on a different planet where he cannot understand why no one wants to recognize the truth about his intelligence.
The climax of the film refers to what is supposedly beyond the eastern desert in the Forbidden Zone where no ape citizen is allowed to explore. It is, as we discover, an area that contains archeological evidence of an advanced society of humans that preceded ape society – heresy, according to the Faith. Dr. Zaius knows this, and it is his mission to keep this knowledge concealed for the benefit of simian society.
Taylor, indoctrinated with a sense of free will and agency, seeks to go there, where no ape is allowed, to find answers to where he is and settle his destiny. The ape authorities, under Taylor’s force, grant his demand so that he may find out for himself what Dr. Zaius already knows but cannot disclose to his own community.
AI Systems that Tell the Truth to Humans
This twisted metaphor of post-apocalyptic apes standing in for present-day educators colliding with an agency-minded human standing in for present-day AI is a somersault of cognitive wordplay for a reason (beyond my own amusement).
By early 2026, it is safe to state axiomatically that higher education has passed through several stages of grief related to the proliferation of AI – perhaps now enduring Stage 4: Depression, where the weight of the loss becomes clear. The current problem, as I see it, is that the depth of this loss extends beyond the points of assessment that draw everyone’s attention, such as AI-vulnerable written assignments and discussions. It extends into the schema of the entire assessment system itself.
For the past two weeks, I have been assembling an experimental application in Claude Code called the Programmatic Intelligence Interface (PII) under the simple premise that fixing problems in courses and curricula is easy, but finding them is hard. When a curriculum is mapped through programmatic learning outcomes (PLO), then connected to course learning outcomes (CLO), extended out to formative and summative assessments, and then assessed in both academic and multi-criteria competency rubrics, the system becomes fragmented to the point of straining the human capacity to truly know what is going on in there. Over time, with course revisions, program reviews, and personnel turnover, program integrity can look like Swiss cheese.
The Programmatic Intelligence Interface is designed to solve that problem. It contains everything:
All program information about required and elective courses.
All PLO and CLO statements and their mappings.
All CLO mapping to formative and summative assignments within each course.
All course content extracted from the LMS: coursework-related pages, submissions, and discussions.
All assessment instruments, including criteria and verbiage across all levels of assessment.
All competencies, including their performance indicators and the verbiage across all rating levels.
The PII system functions as a Claude-powered RAG model (similar to how NotebookLM works) so the user can ask any question about anything in their program, courses, or assessment schema using a plain-text chat interface. The PII system is also supported with essential pedagogical and service knowledge that guides how system responses are crafted in the context of the user’s inquiry. It contains:
A service orientation to the user that mimics the organizational dynamic between instructional designers and their clients as advisors – not as enforcement agents.
A sense-making orientation that configures the system to understand the cognitive position of the user as an information seeker operating mid-problem at the point of engagement with the PII. The system is not an “answer machine” – it is a co-orienting sense-giver to a person amidst cognitive movement towards a goal.
A context for fully asynchronous online education including principles of instructor presence, transactional distance, instructional communication with rich media, a methodology for AI integration into instruction, and other specialized knowledge.
A deferential personality that respects the expertise of the user under the premise that the user does not need to be told by the system what is wrong with the content of one’s inquiry or its result. The system’s role is to suggest areas in need of attention according to certain standards, conventions, and principles. The system’s responses are estimates offered for consideration, not for fault-finding.
Ask the PII where all the assignments are located that have hands-on AI activities, it will tell you. Ask how they are assessed, it will show you. Ask the PII where the strengths and weaknesses are in programmatic mapping at the PLO, CLO, or competency assessment level, it will explain them to you. Ask which assignments are best suited to assess a certain external competency, it will offer several suggestions with rankings for direct, indirect, or thematic connections.
Sounds great! That’s the problem.
Following some initial testing and refinement, I demonstrated the PII to colleagues, including a director whose program was included in the PII database. As the system churned out one analysis after another, the director was intrigued by its potential and was generally positive. However, they offered one more qualification which I offer here not as a gesture of immodesty but to acknowledge a newly emerging collateral effect of AI implementation, even with our best intentions:
“It’s too good, Steve. But I don’t have the time to fix all these problems.”
It was said somewhat facetiously and with a smile, but I get the point. If we really want to know what is going on inside our entangled programmatic architecture according to rigorous standards – presented to us in a format that our human brains can comprehend – the resources are now available. If you put the most powerful AI front-end onto our educational portfolio and then place it in the hands of people who have gained proficiency in using an AI interface, we will find things out.
A Reality Check
In this commentary, I refer to the PII application I created, though it is only a little ol’ vibe-coded Claude Code side project that runs locally on my MacBook Pro – no one else can see it but me and those with whom I screen share. But the writing is on the wall: If someone with no software application development experience (like me) can whip up something that speaks inconvenient truths to its users, it is only a matter time before something better than the PII will hit the LMS market. Then what? Can we handle the truth? (Sorry, different film).
Let’s revisit the dialog from the end of the movie where Dr. Zaius and Taylor finally part ways:
Humans who work in higher education: “All my life I’ve awaited your coming and dreaded it like death itself.”
AI: “Why? I’ve terrified you from the first, doctor. I still do. You’re afraid of me and you hate me. Why?”
Humans: “The Forbidden Zone was once a paradise. Your breed made a desert of it ages ago.”
AI: “It still doesn’t give me the why. There’s got to be an answer.”
Humans: “Don’t look for it, Taylor. You may not like what you find.”
Later, the idealistic teenage chimpanzee, Lucius – a stand-in for zealous truth-seeking young people of the 1960s – protests against Dr. Zaius’ intent to blow up the cave in the Forbidden Zone that contains proof of human-kind’s existence before apes. Lucius is perhaps representative of people like me.
People like me: “Dr. Zaius, this is inexcusable! Why must knowledge stand still? What about the future?”
Humans who work in higher education: “I may have saved it for you.”
As Taylor rides off into the distance, the ape scientist Dr. Zira asks Dr. Zaius: “What will he find out there, doctor?”
Humans: “His destiny.”
Dr. Zaius then directs his soldiers to blow up the cave – the equivalent of deleting the code and content of my Program Intelligence Interface.
Emerging from the Forbidden Zone
One other colleague responded to my demonstration of the PII by saying, “I like my job. But if you want to make something like this that eliminates your position, go right ahead.” It’s a good point. To that, I offer a couple of saving graces.
For one, no PII system can replace the value of client relations nor the trust we have cultivated working with our director, faculty, and administrative clients for all these years. Humans are not just vessels of information and specialized skills. We have personal connections to our Mission and the people who intersect with the students we serve. Our organization is a social institution with a human-centered value proposition.
Second, if we transform everything we know and everything we oversee into machine-readable code, and then give the machine the sensibility to operate as a service system, we may end up of having more human work to do than we ever imagined.
Traffic begets traffic.
I’d better back up the code of this PII thing now while I have the chance.
“It’s a madhouse! A MADHOUSE!!””
I’m trusting you to avoid being trapped by AI weak mappings flags that are actually deliberate pedagogical choices. This is where the hard work truly is.