The AI reasons. The faculty member decides.
The faculty-of-record posture is the central design constraint of ArthurAI™ University Learning Edition. Every artifact the AI generates routes through an explicit review-and-attest checkpoint before it reaches a student or a grade. This is what makes ULE deployable into university programs that care about academic integrity, accreditation, and the faculty member's professional autonomy.
Why the faculty-of-record posture matters
Universities are not classrooms. The faculty member is the certifying-academic-authority for the courses they teach. Their syllabus is theirs. Their grades are theirs. Their advising recommendations are theirs. Accreditation agencies (HLC, MSCHE, SACSCOC, WSCUC, NEASC) review faculty work as the work of the institution. AI-generated content masquerading as faculty work would be a problem at every level — academic-integrity policy, accreditation review, faculty-handbook compliance, intellectual- property attribution. The faculty-of-record posture is how ArthurAI™ avoids that problem by construction.
Where AI shows up in the workflow
- Curriculum. Course-syllabus drafts, weekly-topic outlines, draft assessment plans. The faculty member reviews and revises before attestation.
- Lesson content. The 6-step lesson body for each session. The faculty member reviews before publishing to enrolled students.
- AI tutor. Real-time tutoring for enrolled students with citation discipline pointing back to the lesson source. The faculty member sees aggregate signals (which lesson scopes generated the most tutor questions) but never the conversation content.
- Advising. The platform surfaces students who haven't completed the midterm, students whose engagement has dropped, students whose competency progress lags the cohort. The platform never autonomously decides about progression. The faculty member or academic advisor decides.
- Assessment authoring. AI generates a candidate item pool tied to the course's stated learning outcomes. The faculty member curates the actual assessment from the candidate pool.
- Assessment scoring. Objective items autoscore against the key the faculty member approved. Subjective items get an AI-suggested rubric-aligned score with citations to the student's text. The faculty member reviews, accepts or overrides, and attests.
The review-and-attest checkpoint
Every AI-suggested artifact that affects the academic record reaches the faculty member with three things attached:
- Confidence calibration. Where the AI hedges, the AI flags it. Low-confidence rubric assessments are surfaced for closer review.
- Source citations. The AI shows where it found evidence. For lesson generation, citations point at the curriculum source. For rubric-aligned scoring, citations point at the student's text.
- An attest action. The faculty member must explicitly accept (with or without edits) or override the AI suggestion. The platform does not silently accept on inaction.
The audit trail
Every attestation event is logged with actor identity, timestamp, the artifact, the rubric reference where applicable, and the decision (accepted or overridden, with comment). For accreditation review and academic-affairs audit, the trail demonstrates that the academic record reflects faculty judgment at every node. The platform makes the AI-use posture visible without making it intrusive.
Accreditation alignment
ArthurAI™ provides a documented AI-use posture at institutional onboarding for the academic-affairs office and accreditation liaison. The posture covers the platform's role (decision support), the faculty-of-record framing, the audit trail, the data-handling structure, and the disclosure surfaces students see. We do not represent ArthurAI as a tool that replaces faculty work; we represent it as a tool faculty use to do their work more efficiently. Accreditation agencies have responded well to this framing in early conversations.
What the AI never does
- The AI never autonomously publishes a lesson to enrolled students.
- The AI never autonomously enters a grade in the gradebook.
- The AI never autonomously decides about academic progression.
- The AI never autonomously sends advising communication to a student.
- The AI never autonomously revises a syllabus after attestation.
- The AI never represents itself as the faculty member or as a licensed educator. It represents itself as the ArthurAI tutoring assistant, with the educator-decides motto in every disclosure surface.