A five-course sequence for educators
The Language
Detective.
AI Literacy, Language, and Liability
Five courses, one module each · 2 hours 30 minutes each · 12.5 contact hours
Scope is shaped to the cohort. Courses can be taken as the full sequence, resequenced, combined, or run singly, at the length the setting allows.
Premise
A big-data economy is entrenched in the vendor-district relationship. That economy runs on data. The data arrives as language, the language determines the liability, and a named human has to read it. That read is not only how a school governs its use of AI but how each of us governs our own, because being literate is not enough; the point of reading the language is to choose how we engage with the tools that require and collect our inputs.
This sequence teaches educators to perform that read, to govern their own use of AI, and to carry the read to their students and the families they serve.
Course outcomes
By the end of the full sequence, a participant can do each of the following on a tool already in use in their own setting. A shorter booking carries the outcomes of the courses actually run.
- Explain the data economy beneath school AI, and why student data carries heightened, durable, regulated stakes.
- Locate the AI inside an embedded tool by reading its vendor, product, and data language.
- Apply the SURFACE features to read an output or a vendor claim for where the language outruns the record.
- Author and audit input as both a quality practice and a data-stewardship practice.
- Analyze the author, reader, and subject positions to find the equity exposure in a determination made about a student.
- Use the read to govern their own engagement, and build a classroom, family, or policy artifact, and defend it as the accountable reader.
The running deliverable
Each participant chooses, at the start, one AI tool already in use in their setting, and reads it across the sequence. Each course produces one piece of a single portfolio.
Course
Portfolio increment
ONE
Data-economy profile of the chosen tool
TWO
READ memo: where the AI is, and what it does with what it takes
THREE
OUTWARD findings sheet: named features in its outputs and claims
FOUR
INWARD analysis: a crafted-input study and a configuration-and-data audit
FIVE
BUILD artifact and defense: the read assembled, plus a lesson, family brief, or policy proposal
The completed portfolio is a forensic read of one tool plus one BUILD artifact, defended in the final session.
Session rhythm
Every course runs the same shape, so the method becomes the habit. The proportions hold when a course is shortened; the parts are compressed, not dropped.
Open
15 MIN
A live read. A short specimen on the board, read cold.
Frame
30 TO 35 MIN
The stage. The course’s grounding and its named moves.
Lab
55 TO 60 MIN
Guided practice. The read applied to the chosen tool, in pairs.
Share
30 MIN
Name the move. Findings to the room, each one named, not merely felt.
Bridge
15 MIN
The increment. The course’s portfolio piece and the reading.
The five courses
The courses carry the four stages of the Forensic Read™. Course One lays the ground the stages run on; READ and TRACE open the document stack; SURFACE turns outward on the tool and then inward on the input; BUILD closes with an artifact a person signs.
COURSE ONEGround for all four stages
The Ground: The Data Economy and the Language of Liability
Name what the data economy is and what student data can be used for. State the draft-versus-determination principle: every AI output is a draft built to read as a determination of the matter, when it is only a determination of its input. Frame the read as both school governance and self-governance, the two governances that run through the rest of the course. Participants choose the tool they will read across the sequence and build its data-economy profile from the tool’s own policy: what it collects, what it retains, what it shares, and what it claims to do with it.
Increment 1. Data-economy profile of the chosen tool.
COURSE TWOStage: READ, opening TRACE
READ: Locating the AI in the Product’s Language
Most school AI is invisible. It shows only a number, a level, or a flag, and the only words are the vendor’s. READ answers two questions: where the AI lies, in both senses, where it sits in the product and where it overreaches; and what the tool does with what it takes. Because those answers live scattered across the how-it-works page, the feature list, the settings, the privacy policy, and the marketing, the course also opens TRACE: reading the stack as one ecosystem rather than one page. The light literacy READ set against the full governing READ.
Increment 2. READ memo: where the AI is, and what it does with what it takes.
COURSE THREEStage: SURFACE, closing TRACE
SURFACE Outward: Reading the Output and the Claim
Name the specific language feature carrying a finding, rather than calling an output biased or wrong. The named features, each shown in prose and in interface: presupposition (the guess as settled fact), deferred specification (“research-validated,” nothing named), the self-referential epistemic marker (the tool vouching for itself), nominalization (a bare level number, the guesser deleted), and documentary inconsistency (the policy that contradicts the marketing). That last one closes TRACE: the finding exists only across documents, never inside one. Bias rides in the wording before it rides in the consequence.
Increment 3. OUTWARD findings sheet: named features in the tool’s outputs and claims.
COURSE FOURStage: SURFACE, turned on the input
SURFACE Inward: Authoring and Auditing the Input
The inward read in two halves. Craft: the draft is bounded by the input’s quality, and intentional input is also data stewardship, controlled, informed, and intentional, which is release only what the task needs, read what the tool does with it, and author the rest on purpose. Check: the planted conclusion, the silent setting, the wanted answer you decline to challenge. For a tool you do not prompt, the input is a student’s behavior plus a configuration, and the audit falls to the adult who set it.
Increment 4. INWARD analysis: a crafted-input study paired with a configuration-and-data audit.
COURSE FIVEStage: BUILD
The Positions and the Build: Equity, Accountability, and the Cascade
The author, reader, and subject positions, and how they split for student-facing tools: the child is the subject who never reads the determination, and the most-sorted students are the ones most often made subjects. The accountable reader is the equity intervention. BUILD as register and genre construction, and self-governance as the language to bring to students and parents. Each participant assembles the read into one account of the tool, produces one BUILD artifact, then stands behind it and takes questions from the room.
Increment 5, final. BUILD artifact and defense. With the four prior increments, this completes the portfolio.
What stands alone, and what does not
Courses One, Two, and Three
Stand on their own. Each can be run as a single session for a cohort that has never met the method, and each leaves participants with a move they can perform the next day.
Course Four depends on Course Three
The inward check is performed with the named features as its vocabulary. Without them a participant can feel that an input was leading, but cannot say what the language did, which is the whole point of naming the move.
Course Five depends on the read it assembles
Run on its own it delivers the positions and the equity analysis, and that much holds. The BUILD artifact and the defense need the prior increments to stand on, because a defense with nothing behind it is a discussion.
Assessment and the cascade
The course is assessed on the single accumulating portfolio and its defense.
How a participant completes it
The portfolio, five increments. A full forensic read of one AI tool in the participant’s setting: its data economy, its located AI, its outward and inward reads, and its equity exposure, plus one derived BUILD artifact.
The final defense. The participant stands behind the read as the named, accountable reader. This is the firm’s primary principle made into the final exam: accountability for what a tool does belongs to a person, not the tool.
The cascade is the output
Because each BUILD artifact is a lesson, a family brief, or a policy proposal, the course does not only train forensic readers. It produces the student-facing and parent-facing versions of the read as its own product.
A building that runs a cohort finishes holding classroom lessons, family communications, and policy language it did not have before, each one authored by a named person who can defend it.
Included with every booking
No technical background is assumed. The foundation is supplied in writing before the first session.
The Foundational AI Primer
A plain-language foundation: what generative AI is, how a language model produces its output, what it cannot reliably do, and the vocabulary that recurs across the tools and their terms.
What is the tool actually doing? Name the AI-driven function, not the marketing. A generator, a tutor, a grader, and a predictor carry different risks.
Who is it for? A student-facing tool and a teacher-facing tool place the same technology in very different hands, and call for different supervision.
What becomes of the data? What is entered into a tool is governed by the vendor’s terms, not by the tool’s manner. Treat silence in those terms as the most permissive reading available to the vendor.
Scope
Stated plainly, so whoever books the cohort knows what arrives and what does not.
Who it is for, and how it runs
- Classroom teachers, instructional and technology coaches, librarians and media specialists, department leads
- Cohorts are kept small and work in pairs. The read is learned by doing it on a real tool, not by lecture
- Every participant brings one tool from their own setting and reads it across the sequence
- A single worked example runs through every course: the same read performed on three public documents, which participants then perform on a chatbot answer, a product page, and a flagging screen
What it is not
It is not a compliance sign-off. The course trains readers; it does not clear a tool, approve a purchase, or produce a district audit.
It is not legal advice. The firm reads language and documents what the language does. Counsel advises on liability.
It is not technical training, and it endorses or ranks no tool. It supplies the working concepts an educator needs to judge a tool and to supervise its use.
How a booking runs
Scoped before anything is invoiced, and confirmed in writing before any work begins.
STEP IRequest and scope
You describe the cohort: who attends, which tools are already in use, how much time the calendar allows, and what the group needs to be able to do afterward. The firm proposes which courses to run, in what order, and at what length. Nothing about the sequence is fixed before this conversation.
↓
STEP IISpecimens assembled
Sessions run on real documents. The firm assembles the specimen set for the tools your building actually uses, so participants read the language they are already governed by. The Primer goes to every participant ahead of the first course.
↓
STEP IIIConfirmation in writing
You receive written confirmation of the courses to be run, their length, fees, and dates within 24 hours of the scoping conversation. Work begins on a signed confirmation or purchase order.
Rates
The full sequence is the complete product. Anything shorter is a compressed consult drawn from the same five courses, scoped to what the cohort needs and the calendar allows, without the portfolio and the defense.
The full sequence · five courses
All five courses at 2 hours 30 minutes each. 12.5 contact hours. Includes the running portfolio, the final defense, the assembled specimen set, and the Primer for every participant. Scheduled weekly, or condensed or spread to fit the calendar.
Compressed consult
Individual courses of 60 to 90 minutes, chosen and sequenced with the cohort in mind. No portfolio, no defense, no contact-hour package.
One courseSixty minutes minimum$500.00
Two coursesSixty minutes or more each$1,500.00
Three coursesSixty minutes or more each$2,500.00
Four coursesSixty minutes or more each$3,500.00
Extended, repeated, or multi-cohort scopeScoped in conversationOn request
The full sequence produces 12.5 documented contact hours plus between-course work, which a host program or district can map to its own continuing-education or graduate-credit requirements, subject to its verification. Extended courses, added cohorts, and custom sequences are quoted in the scoping conversation. Final scope, fees, and dates are confirmed in writing within 24 hours, before any work begins.