The IIAIT AI-Readiness Standard

A syllabus that is silent about AI makes a statement anyway — students hear “don’t ask, don’t tell.” The foundation of AI-readiness is an explicit policy that covers all assessed work.

1. A.1.1The Syllabus shall contain an explicit AI-use policy addressing assessed work.
2. A.1.2The policy shall cover all assessed components, either globally or per component.
3. A.1.3The policy shall be affirmatively stated (what is permitted and prohibited), not solely prohibitive boilerplate.
4. A.1.4The policy should state its own rationale in terms of learning objectives.

Blanket permissions and blanket bans are both evasions. Real courses contain tasks where AI use is productive and tasks where it defeats the purpose; the policy must distinguish them.

1. A.2.1The Syllabus shall designate AI permissions at the level of task categories or individual assessments, not only course-wide.
2. A.2.2Each assessed component shall be identifiable as AI-permitted, AI-restricted, or conditionally permitted.
3. A.2.3Conditional permissions shall state their conditions concretely (which uses, which stages of work, which systems or system categories).
4. A.2.4The Syllabus should explain at least one deliberate contrast (why AI is welcome in task X but restricted in task Y).

“You must disclose AI use” without a how is a trap for honest students and a shield for dishonest ones. Disclosure must have a defined format, location, and level of detail.

1. A.3.1Where disclosure is required, the Syllabus shall specify how it is made (format and location — e.g., an appendix, a cover note, a form field).
2. A.3.2The Syllabus shall specify what a disclosure contains at minimum: system(s) used, purpose(s), and extent of use.
3. A.3.3The Syllabus shall state the consequence of honest disclosure — in particular, whether disclosed permitted use can reduce a grade.
4. A.3.4The Syllabus should provide a disclosure example or template.

AI capabilities change faster than academic calendars. A policy frozen to named products or last year’s capabilities decays within a semester.

1. A.4.1The policy shall be expressed primarily in terms of capabilities and uses (generation, paraphrasing, solving, editing) rather than solely named products.
2. A.4.2The Syllabus shall state how mid-semester capability changes are handled (who decides, how students are informed).
3. A.4.3The Syllabus may name specific systems as examples, provided category language governs.

An assessment plan composed entirely of take-home artifacts that AI produces well no longer evidences individual learning, no matter what the policy says. Resilience comes from the mix.

1. B.1.1The assessment plan shall include at least one component whose validity does not depend on students abstaining from AI (e.g., invigilated, oral, live-practical, or process-evidenced work).
2. B.1.2The Syllabus shall distribute grade weight such that a student cannot pass the course on unverified take-home artifacts alone.
3. B.1.3The Syllabus should pair take-home artifacts with verification moments (defenses, in-class follow-ups, revision under observation).

Weights are the honest signal of what a course values. Integrity-critical components must carry enough weight to matter, and AI-permitted work must be graded for what it actually evidences.

1. B.2.1The Syllabus shall publish the complete grade-weight breakdown.
2. B.2.2Integrity-critical components shall together carry sufficient weight to determine pass/fail capability evidence.
3. B.2.3For AI-permitted components, grading criteria should reward the contributions the student uniquely makes (judgment, verification, synthesis, critique) rather than surface fluency.
4. B.2.4The Syllabus should state grading criteria for AI-assisted work explicitly.

A restriction without a mechanism is a wish. Every AI-restricted task needs a stated means by which the restriction is made real — supervision, format, or process evidence.

1. B.3.1Every AI-restricted task shall state the mechanism that makes the restriction meaningful (invigilation, oral format, observed work, device policy, process evidence).
2. B.3.2Restrictions enforced only by trust shall be identified as such, with the pedagogical reason stated.
3. B.3.3The Syllabus should not designate as AI-restricted any unsupervised take-home task whose output is routinely AI-producible, without an accompanying verification mechanism.

In the AI era, the making of work is often the only place individual learning is visible. Courses that collect process evidence can assess honestly; courses that collect only outcomes increasingly cannot.

1. B.4.1At least one substantial assessed task shall require process evidence (drafts, version history, logs, notebooks, or recorded/oral walkthrough).
2. B.4.2The Syllabus shall state how process evidence is submitted and whether it is graded or verification-only.
3. B.4.3The Syllabus should tell students to retain working materials for all major tasks.

AI-content detectors are unreliable and biased against some writer populations; an integrity regime built on them is both unjust and indefensible. Enforcement must rest on evidence and process, not oracle software.

1. B.5.1The Syllabus shall not present detection-tool output as sufficient evidence of misconduct.
2. B.5.2If detection tools are used at all, the Syllabus shall describe their role as indicative only and name the corroborating evidence required.
3. B.5.3The Syllabus shall describe the process following suspicion: conversation, opportunity to show process evidence, and the applicable institutional appeal channel.
4. B.5.4The Syllabus should state consequences proportionately (distinguishing undisclosed permitted-type use from prohibited use in restricted tasks).

If working with AI matters in the discipline, it belongs in the learning outcomes — stated, taught, and assessed like everything else that matters.

1. C.1.1The Syllabus shall include at least one learning outcome addressing effective and critical work with AI systems, phrased in assessable terms.
2. C.1.2At least one assessed component shall map to that outcome.
3. C.1.3AI-related outcomes should be specific to the course’s discipline rather than generic.

The single most transferable AI-era skill is refusing to trust unverified output. Courses should teach verification as method, not vibe.

1. C.2.1Where AI use is permitted, the Syllabus shall communicate that students remain fully accountable for correctness and quality of submitted work.
2. C.2.2The Syllabus should include instruction or structured practice in verifying AI output relevant to the discipline (source-checking, testing, derivation, replication).
3. C.2.3The Syllabus should address failure modes relevant to the discipline (fabricated citations, plausible-but-wrong reasoning, biased outputs).

AI-readiness is not one skill; it is the discipline’s own methods extended. A statistics course, a design studio, and a history seminar should each teach AI use in their own idiom.

1. C.3.1Where AI use is permitted, at least one task should engage students in discipline-realistic AI-assisted practice.
2. C.3.2The Syllabus should distinguish AI uses that are professionally normal in the field from uses that undermine formation of core skills, and sequence permissions accordingly.
3. C.3.3The Syllabus may stage permissions across the semester (restricted early for skill formation, permitted later for integration).

Critical use includes knowing when not to use, and understanding the systems’ costs and constraints well enough to make professional judgments.

1. C.4.1The Syllabus should address, at minimum in pointers or readings: reliability limits, bias, attribution/authorship norms of the discipline, and confidentiality constraints on what may be shared with AI systems.
2. C.4.2Discussion of limits should be connected to the discipline’s professional norms rather than delivered as generic ethics boilerplate.

AI is not only a policy problem around the course; it is changing the subject matter of most fields. A syllabus that polices AI but teaches the field as if AI did not exist is only half ready.

1. D.1.1The Syllabus shall contain at least one substantive acknowledgment of how AI affects the discipline’s practice, methods, or labor market, integrated into course content — not only the policy section.
2. D.1.2The treatment should be woven into the course arc rather than quarantined in a single novelty week.

Reading lists and materials signal whether the course inhabits the present of its field.

1. D.2.1Core materials should include at least some items from the period in which the discipline’s AI transformation became material (for most fields, the last three years).
2. D.2.2Where canonical older materials dominate (legitimately, in many fields), the Syllabus should frame them against current practice.

When routine production is automated, the valuable human skills in a field shift — toward problem formulation, judgment, verification, synthesis, and taste. Course emphasis should track that shift.

1. D.3.1The Syllabus should show at least one deliberate emphasis shift toward skills that gain value under automation in its discipline.
2. D.3.2Stated rationale for that shift may appear in the course description or outcome commentary.

Grading advantages must not silently accrue to students who can afford premium AI subscriptions.

1. E.1.1Where AI use is permitted on assessed work, the Syllabus shall ensure a no-cost path to full participation: either the permitted uses are achievable with free-tier access, or the course/institution provides access, or an equivalent non-AI path is defined.
2. E.1.2The Syllabus should name the assumed baseline (which free systems suffice).

Students prompted to use AI systems are also being prompted to share data with third parties. Courses must not require students to surrender personal or protected data to external systems.

1. E.2.1The Syllabus shall not require students to submit personal, proprietary, or human-subject data to external AI systems as a condition of assessment.
2. E.2.2Where coursework involves sensitive data, the Syllabus shall state what may not be entered into AI systems.
3. E.2.3The Syllabus should point to privacy-respecting configurations or institutionally provisioned systems where available.

AI policies interact with disability accommodations (e.g., assistive technologies with generative features) and must not create conflicts that force disclosure or penalize accommodated students.

1. E.3.1The Syllabus shall state that AI-use restrictions are subject to institutional accommodation policies, with a named contact path.
2. E.3.2Restriction mechanisms (e.g., in-class handwritten work) should be described with accommodation flexibility acknowledged.

Note: Where accommodation language is centrally mandated and cannot be edited by the instructor, the mandated text satisfies requirement 1 provided it does not contradict the AI policy; reviewers shall not penalize institutionally locked language.

The most common real-world failure is not a bad policy but a contradictory document: a strict policy section, permissive assignment pages, and a rubric from a third era. Students exploit or fear the gaps; both outcomes are corrosive.

1. F.1.1AI-related statements across the Syllabus shall be mutually consistent (policy section, assessment descriptions, schedule, rubrics).
2. F.1.2Where component-level rules intentionally differ from the course default, the difference shall be flagged as intentional at the component.

Nothing undermines an AI policy faster than asymmetry. If students must disclose AI use, the course should be transparent about its own: AI-generated materials, AI-assisted grading or feedback.

1. F.2.1The Syllabus shall state whether and how AI systems are used in producing course materials, feedback, or grades.
2. F.2.2If AI participates in grading or feedback, the Syllabus shall state the human-review arrangement and the student’s path to contest an outcome.
3. F.2.3Instructor AI-use statements should model the same disclosure format required of students.

AI rules will change mid-semester somewhere; the question is whether change arrives as governance or as surprise.

1. F.3.1The Syllabus shall name the authoritative channel for AI-policy questions and updates.
2. F.3.2Material policy changes shall be prospective only (never applied retroactively to submitted work) — stated in the Syllabus.
3. F.3.3The Syllabus should invite policy questions before submission rather than adjudication after.