The emergence of generative artificial intelligence has fundamentally transformed how we approach assessment in higher education. While various frameworks have been proposed to address these challenges, two complementary approaches—the two-lane approach and the AI Assessment Scale (AIAS)—offer a potent combination when used together. This blog post examines how these frameworks can work in synergy to create more transparent, equitable, and pedagogically sound assessment practices. Usually they are set in competition to each other!
Understanding the foundations
The two-lane approach, developed by Danny Liu and Adam Bridgeman from the University of Sydney, conceptualises assessment through two distinct pathways: supervised assessments (where AI use is either prohibited or strictly controlled) and unsupervised assessments (where students have greater flexibility in AI engagement). This approach emerged from the recognition that simply creating “AI-proof” assessments is neither feasible nor desirable in preparing students for a world where AI is ubiquitous.
The AI Assessment Scale (AIAS), developed by Mike Perkins, Leon Furze, Jasper Roe, and Jason MacVaugh, provides a nuanced framework for integrating AI into assessment across five levels:
- Level 1 (No AI),
- Level 2 (AI Planning),
- Level 3 (AI Collaboration),
- Level 4 (Full AI), and
- Level 5 (AI Exploration).
Notably, the authors emphasise that:
“we see the AIAS primarily as an assessment design tool, not an assessment security tool. For the reasons explained elsewhere in this commentary, it is neither possible nor fair to simply label a task at a given level of the AIAS and then expect students to comply with it. If you permit GenAI, redesign the brief, evidence trail, and rubric to grade students’ decisions, checks, and justifications, not just their ability to push buttons”[Perkins et.al, 2025, p.4].
Previously, many viewed the AIAS primarily through the lens of assessment security—a way to control and monitor AI usage. However, this perspective fundamentally misunderstands the framework’s true potential. The AIAS is not about enforcing compliance but about redesigning assessment practices to align with learning outcomes in an AI-enhanced world.
When combined, the supervised lane of the two-lane approach handles assessment security through controlled conditions, freeing the AIAS to function as its creators intended: a design tool for authentic, pedagogically meaningful assessments.
The synergistic relationship
The integration of these frameworks creates a powerful synergy. The closed lane (supervised assessments) of the two-lane approach satisfies the need for AI-free validation points where students must demonstrate independent mastery of core disciplinary knowledge. This addresses legitimate concerns about academic integrity while ensuring qualifications retain their credibility.
Simultaneously, the open lane (unsupervised assessments) becomes a space for sophisticated AI integration using the AIAS framework. Here, educators can design assessments that span levels 2-5 of the scale, focusing on developing students’ ability to work effectively and ethically with AI tools. This transparency removes the performative aspects of “no AI” declarations in contexts where enforcement is impossible.
Practical applications across AIAS Levels
Level 1 (No AI) – The Supervised Lane
In the supervised lane, Level 1 assessments serve as crucial validation points. Examples include:
- In-class analytical essays where students demonstrate independent critical thinking
- Oral examinations that require real-time articulation of disciplinary knowledge
- Supervised laboratory practicals where students must demonstrate hands-on competencies
- Timed problem-solving sessions that assess foundational mathematical or logical reasoning skills
These assessments provide what Dawson (2020) identifies as essential components of assessment security: authentication and control of circumstances.
Level 2 (AI Planning) – The Open Lane
In the unsupervised lane, Level 2 assessments allow students to use AI for initial research and planning while maintaining human agency in analysis and synthesis:
- Research proposal development where students use AI tools like Elicit.org to identify relevant literature, but must critically evaluate and synthesise findings independently
- Project planning documents that incorporate AI-generated timelines or resource lists with student justification for choices
- Annotated bibliographies where AI assists in source discovery, but students provide critical analysis and relevance assessment
Level 3 (AI Collaboration) – The Open Lane
Level 3 assessments embrace AI as a collaborative partner while requiring students to demonstrate critical evaluation and decision-making:
- Collaborative research reports where students engage AI as a research assistant, but must document their interactions and justify their acceptance or rejection of AI suggestions
- Creative writing projects that incorporate AI-generated elements (characters, plot points, or dialogue) with student reflection on the integration process
- Data analysis projects where AI tools assist with visualization or pattern identification, but students must interpret findings and draw conclusions
Level 4 (Full AI) – The Open Lane
Level 4 assessments acknowledge that some professional contexts require seamless human-AI collaboration:
- Business case studies where students use AI extensively for market research, financial modelling, and presentation creation while being assessed on strategic decision-making and justification
- Digital humanities projects that leverage AI for text analysis, translation, or multimedia creation, with evaluation based on project design and critical interpretation of AI outputs
- Policy analysis papers where AI assists with research, drafting, and editing, but the assessment focuses on argument quality, evidence evaluation, and ethical considerations
Level 5 (AI Exploration) – The Open Lane
Level 5 assessments push boundaries by encouraging students to explore AI’s limits and possibilities:
- Experimental design projects where students investigate AI capabilities and limitations within their discipline
- Critical AI literacy portfolios documenting students’ evolving understanding of AI’s role in their field through experimentation and reflection
- Innovation challenges where students develop novel applications of AI tools for disciplinary problems with assessment based on creativity, feasibility, and ethical considerations
Benefits of the Combined Framework
This integrated approach offers several significant advantages:
- Transparency and trust: Students understand precisely what is expected in each assessment context. The supervised lane provides clear boundaries, while the open lane offers explicit guidelines for AI engagement.
- Pedagogical coherence: The framework aligns with educational goals rather than working against them. Students develop both independent disciplinary competencies and AI literacy skills essential for their future careers and professional endeavours.
- Equity and accessibility: By providing free access to required AI tools in open lane assessments and ensuring supervised assessments don’t disadvantage students who lack resources, the combined framework addresses equity concerns that plague purely restrictive approaches.
- Authentic assessment: Assessments reflect real-world contexts where AI tools are available and expected to be used ethically and effectively.
- Reduced academic integrity violations: Clear expectations and appropriate task design minimise the likelihood of inappropriate AI use while teaching students how to engage ethically with these tools.
Implementation Considerations
Successful implementation of this combined framework requires careful attention to several factors:
- Program-level design: Both frameworks emphasise the importance of coherent assessment strategies across entire programs rather than ad-hoc changes to individual assessments.
- Staff development: Educators require training in both frameworks, including understanding how to design valid assessments for each AIAS level and implementing effective supervised and unsupervised assessment strategies.
- Clear communication: Students must understand the rationale behind different assessment approaches and receive explicit guidance on expectations for each context.
- Regular review: As AI technologies continue to evolve rapidly, assessment strategies must be regularly reviewed and updated to maintain their relevance and effectiveness.
Conclusion
The combination of the two-lane approach and the AI Assessment Scale offers a comprehensive framework for navigating assessment in the age of AI. By recognising the AIAS as a design tool rather than a security measure and leveraging the two-lane approach to provide both secure validation and authentic AI-integrated learning experiences, educators can create assessment systems that prepare students for their future while maintaining the integrity of their qualifications.
This integrated approach moves beyond the binary thinking that has characterised much of the AI and assessment discourse. Instead of viewing AI as either a threat to be eliminated or a solution to be embraced uncritically, it offers a nuanced framework that acknowledges both the challenges and opportunities AI presents to higher education.
As we continue to adapt our educational practices to an AI-enhanced world, the thoughtful integration of these frameworks provides a roadmap for creating assessment systems that are transparent, equitable, pedagogically sound, and fit for purpose in preparing graduates for their professional futures.
References
Bearman, M., Tai, J., Dawson, P., Boud, D., Ajjawi, R. (2024). Developing evaluative judgement for a time of generative artificial intelligence. Assessment Evaluation in Higher Education, 00, 1–13. https://doi.org/10.1080/02602938.2024.2335321
Bellamy, C. (2024, October 14). Review of (some) debates in AI and Assessment. Craig Bellamy. https://www.craigbellamy.net/2024/10/14/review-of-some-debates-in-ai-and-assessment/
Bellamy, C. (2025, May 19). Reconsidering assessment in higher education: The lane approach. Craig Bellamy. https://www.craigbellamy.net/2025/05/19/reconsidering-assessment-in-higher-education-the-lane-approach/
Perkins, M., Roe, J., & Furze, L. (2025). How (not) to use the AI Assessment Scale. Journal of Applied Learning & Teaching, 8(2), 1-10. https://doi.org/10.37074/jalt.2025.8.2.15
Corbin, T., Bearman, M., Boud, D., Dawson, P. (2025). The wicked problem of AI and assessment. Assessment Evaluation in Higher Education, 00, 1–17. https://doi.org/10.1080/02602938.2025.2553340
AI Statement
This blog post was created with the assistance of artificial intelligence. The author provided the conceptual framework, specific requirements for the 1,000-word length, academic citations in APA 7 format, and detailed guidance on integrating the two-lane approach with the AI Assessment Scale.
The AI tool (Claude) was used to:
- Structure and organise the content according to the author’s specifications
- Research and synthesise information from the provided academic sources
- Format citations and references in APA 7 style
- Generate practical examples for each level of the AIAS framework
- Ensure coherent flow and academic writing style appropriate for a higher education blog
The AI accessed three primary sources as requested by the author: the attached PDF article “How (not) to use the AI Assessment Scale” by Perkins et al. (2025), and two of the author’s previous blog posts on the two-lane approach and AI assessment debates. All conceptual insights, arguments about the complementary nature of these frameworks, and the specific examples provided were generated through AI synthesis of these materials.
The final content reflects the author’s vision for demonstrating how these two assessment frameworks can work synergistically, with AI serving as a collaborative tool in the writing and research process while maintaining the author’s voice and academic perspective throughout.
This statement follows the transparency principles advocated in the blog post itself, demonstrating appropriate disclosure of AI assistance in academic and professional writing contexts.
Leave a Reply