Assessment
The assessment domain modernises assessment design, sets rules for legitimate and illegitimate AI use, and ensures any automated scoring or AI-mediated assessment is subject to bias evaluation and teacher review.
Caribbean context
Continuing to use assessment formats that assume no AI access produces unreliable results and creates equity problems for students with different levels of AI access.
3 provisions in this domain
A Caribbean scenario
A geography department redesigns a coursework task so students must combine AI-generated background research with original local field data, a teacher conferral and a reflective journal. AI is a permitted research tool; the assessed work is unmistakably the student's.
Responsibility matrix
- •Maintain guidance on AI use in assessment
- •Coordinate bias evaluation requirements across endorsed tools
Preconditions for implementation
Where to start
- 01Publish guidance on legitimate AI use in coursework and examinations
- 02Review high-stakes assessment formats for AI-resilience
- 03Require independent bias evaluation for any AI-mediated assessment tool
What progress looks like
- Assessment validity is preserved across AI access levels
- Bias evaluations are completed before tools are used in assessment
- Students understand what AI use is legitimate and what is not
Likely risks and practical responses
MitigationPair AI restrictions with assessment redesign that asks students to do things AI cannot do alone.
What this domain looks like in the roadmap
- RI-D01NowConduct assessment readiness reviewsReview the implications of digital tools and AI for classroom assessment, national examinations, academic integrity, certification, accessibility and equity.
- RI-D02NowEstablish acceptable-use rules for digital and AI tools in assessmentDefine what use of digital and AI tools is permitted, prohibited, supervised, disclosed or required in formative and summative assessment.
- RI-D03NowEstablish academic integrity and authorship guidanceProvide guidance on plagiarism, AI-generated work, disclosure, verification, authorship and appropriate teacher and institutional responses.
- RI-D04NextExpand authentic and resilient assessment approachesIncrease the use of oral assessment, projects, portfolios, practical demonstrations, supervised tasks and other approaches that strengthen validity in AI-enabled learning environments.
- RI-D05NextDevelop AI-era examination reform optionsWork with national examination and assessment bodies (including CXC) to develop options for adapting high-stakes assessment, certification and examination systems.
- RI-D06NextEstablish assessment technology assurance requirementsDefine requirements for accessibility, security, offline functionality, identity verification, data protection, auditability and continuity for digital assessment tools.
- RI-D07LaterPilot validated AI-supported assessment feedback toolsTest AI-assisted feedback tools in controlled settings, with teacher review, learner safeguards and independent evidence of effectiveness.
Jump to another domain or return to the system map.