Clarify the evidence
Identify what the task needs students to know, do, and demonstrate independently or with AI.
AI Assessment Scale
Redesign assessment for a world where AI cannot be assumed away.
The AIAS gives educators five practical design patterns for connecting learning outcomes, valid evidence, student judgement, and the role of generative AI. Communication matters, but redesign is the point.
2023First introduced
5Levels of AI use
30+Language translations
OpenNon-commercial resources
Purpose
A framework for redesigning valid evidence of learning.
The AIAS was originally developed by Mike Perkins, Jasper Roe, Leon Furze and Jason MacVaugh. It provides a shared structure for redesigning assessment around the capability a task should evidence and the conditions under which that evidence remains valid.
It is not intended to enforce every student interaction with AI. Used well, each level changes the task, its conditions, the evidence students produce, the judgement they must demonstrate, or the way that work is assessed.
Learning Innovation Practice supports the continuing stewardship, commercial licensing, and institutional implementation of the AIAS. The public resource collection includes downloads, editable assets, translations, research, and implementation guidance.
AIAS v2.1 is the current version. Each level now has a single statement rather than the earlier pair of educator-facing and student-facing paragraphs, and Levels 2, 3 and 4 state explicitly what is being assessed. The FAQ sets out what changed at each level and why, and the implementation guide covers the practical steps, including the design test for Level 3 and the use of assessment twins to pair open and secured tasks.
Downloads, the editable template, translations and the AIAS Advisor are all on the resources page.
A level attached to an unchanged task is not implementation.
AIAS should help educators redesign assessment briefs, rubrics, evidence, and conditions. The level then gives staff and students a shared language for understanding that design.
Using AIAS well
Use the level as a design instruction, not a permission sticker.
The process begins with the capability and evidence that matter, then works outward to the task, rubric, conditions, and student guidance.
Examine AI's role
Decide where AI supports the intended learning and where it could obscure or replace essential evidence.
Redesign the task
Change the brief, process, conditions, evidence, and rubric so the intended capability remains assessable.
Set and explain the level
Choose the level that matches the redesigned task, then explain what students must produce and demonstrate.
Five design patterns
Each level answers a different assessment design question.
The levels describe what capability the task is designed to evidence. They do not attempt to classify every action a student takes while completing it.
Design controlled evidence
Use supervised or controlled formats, such as examinations, vivas, or practical demonstrations, when unaided capability must be assessed.
Make planning the assessment
Assess planning, synthesis, ideation, storyboarding, or research as the task itself, rather than trying to police a planning-only phase.
Design for visible judgement
Allow substantive AI support while requiring students to evaluate, modify, justify, and take responsibility for the resulting work.
Assess purposeful AI direction
Build the task around how effectively students direct and critically evaluate AI to achieve specified learning outcomes.
Co-design new forms of work
Invite students and educators to use AI creatively to develop methods and outputs that extend disciplinary practice.
Institutional support
Make assessment redesign systematic across programmes
Assessment redesign programmes
Structured audits and hands-on redesign of real tasks, briefs, evidence chains, and marking criteria.
Staff development
Workshops that build shared design capability rather than stopping at awareness or level selection.
Policy and quality alignment
Connect redesigned tasks with assessment regulations, academic integrity, programme quality, and student guidance.
Licensing and resources
Commercial permissions and tailored materials that support implementation in local systems and terminology.
AIAS v2.1
The scale at a glance
Open the image for a closer view, or visit the AIAS website for editable assets, translations, research, and FAQs.
Access and licensing
Public resources for educators. Tailored support for institutions.
Public AIAS resources are available for non-commercial use under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 licence.
Contact Learning Innovation Practice Limited about commercial licensing, commissioned training, tailored implementation, adapted resources, or institutional partnerships.
AIAS enquiries