From the Student Success edition of November 3, 2026
University of Liverpool explains its four-tier AI assessment framework to students, from no AI (Tier 1) to required responsible AI use (Tier 4), shown in Canvas
University of Liverpool · Higher education assessment and academic integrity · United Kingdom (Liverpool)
- Publisher
- University of Liverpool News
- Original publication
- University of Liverpool Student News, 3 November 2026
- Source retrieved
- 2026-11-04
What happened
On 3 November 2026 the University of Liverpool published student guidance on the AI policy it introduced in August. Every assessment is placed in one of four tiers displayed in Canvas: Tier 1 bans all AI including grammar checkers unless an approved student support plan requires it; Tier 2 allows limited contextual support with declared use; Tier 3 has students critique and revise AI drafts with citation; Tier 4 requires responsible AI use. Programmes use a mix, with relatively few requiring AI. All taught students get secure Microsoft Copilot and nebulaONE, a gateway to approved LLMs. Fabricated references or data, deepfakes, false authorship and uploading personal or sensitive data are prohibited in every tier, and professional regulators may impose stricter rules.
Why it matters
A concrete, student-facing model for course-level AI permission that U.S. student-success and academic-integrity offices can compare with their own syllabus-by-syllabus approaches.
Evidence and measured results
Primary institutional guidance; describes policy, platforms and tier definitions; no evaluation data.
Limitations and uncertainty
UK institution; transferability to U.S. institutions depends on local honor codes and accreditation; no outcome or compliance data.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-11-04; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Students face inconsistent, course-by-course AI rules and unequal access to AI tools.
- Stakeholders
- Provost or academic quality office
- Academic integrity office
- Disability services
- Library
- Central IT
- Discovery questions
- How do students learn the AI rules for each assessment today?
- Do all students have access to approved AI tools?
- How are disability accommodations handled when AI is banned?
- Value hypothesis
- Interpretation: a shared tier taxonomy shown in the LMS may reduce student confusion about permitted AI use.
- Potential engagement
- A sprint to design an assessment AI tier taxonomy and surface it in the LMS.
- Unsupported claims
- Tiers reduce academic misconduct
- Tiers improve learning outcomes
- Students comply with declarations
Pre-sales engineering
Role takeaway
- Fit
- LMS assessment metadata plus a secure approved-model gateway.
- Architecture and integration
- Tier label shown in Canvas per assessment; institution Copilot and nebulaONE gateway for AI-required work.
- Prerequisites
- Approved-model list
- Identity integration for the gateway
- Declaration and citation templates
- Constraints
- Professional regulators may set stricter rules
- Assistive-tool exceptions under support plans
- Non-UK legal context differs
- Security
- Prohibit sensitive-data uploads; route approved AI through a gateway with logging and data controls.
- Proof of value
- Share of assessments with a tier label
- Gateway usage on AI-required assessments
- Student survey on rule clarity
Delivery
Role takeaway
- Work
- Assign tiers to assessments, configure LMS labels and train students and staff.
- Dependencies
- Approved AI policy
- LMS configuration
- Library declaration guidance
- Ownership
- Academic quality or teaching and learning office.
- Skills and adoption
- Module leaders need guidance on choosing tiers; students need tutorials on declaring AI use.
- Governance checkpoints
- Policy approval
- Accessibility review of Tier 1 exceptions
- Regulator alignment for professional programmes
- Acceptance criteria
- Proposed: tier label on all assessments
- Proposed: student rule-clarity survey above an agreed threshold
- Proposed: documented support-plan exception process
- Risks
- Inconsistent tier assignment
- Accommodation conflicts
- Over-reliance on declarations
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Tier labels are surfaced in the LMS (Canvas); students use institution-provided Copilot and a secure model gateway (nebulaONE) for AI-required assessments.
Governance
Who approves, reviews and stays accountable for outcomes?
A shared tier vocabulary gives students one rule set across modules, with disability support plans and professional regulators as explicit exceptions.
Security and privacy
What data, permissions and controls need testing?
Uploading personal, confidential or sensitive data to AI is prohibited in all tiers; approved tools are routed through a secure gateway.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Tier 1 explicitly preserves assistive tools required under approved student support plans.
Procurement
What should contracts, pricing and exit terms secure?
Providing approved AI to all taught students reduces inequity from paid consumer tools but requires licensing and gateway contracts.
Operating model
Which teams own the service once it runs?
Module leaders assign tiers; the library supplies declaration guidance; tutors are the first point of contact for questions.
Publication history
- 2026-11-03Student Success · Issue 574 resources
Stable resource ID: liverpool-four-tier-ai-assessment-framework-student-guidance-20261103