Three pathways to assessment reform: what TEQSA’s AI guidance means for universities

AI is now embedded in how learners study, the tools educators use, and the skills employers expect. That reality brings a challenge: assessments have to evolve. They need to maintain trust in qualifications while preparing learners for professional contexts where AI is part of everyday practice.

In Australia, the Tertiary Education Quality and Standards Agency (TEQSA) is the national regulator responsible for safeguarding higher education standards. In September 2025, TEQSA released Enacting assessment reform in a time of artificial intelligence. The resource is both pragmatic and ambitious. It recognises that AI detection tools can’t guarantee integrity and that structural redesign is the only sustainable response.

Although the guidance is aimed at universities, the issues it raises resonate across education. Schools, RTOs, and training providers are grappling with the same question: how do we design assessments that are credible, authentic, and future-facing in the age of AI?

Why TEQSA weighed in

Since generative AI became mainstream in 2022, institutions have taken different paths. Some banned tools outright, others experimented widely, and many focused on detection. TEQSA’s new guidance shifts the emphasis. It highlights that the sector needs systemic change rather than piecemeal fixes.

This follows a significant milestone in July 2024, when TEQSA required every Australian higher education provider to submit an institutional action plan on generative AI. All 203 providers responded, a 100% response rate that underscored both the scale of the challenge and the urgency with which the sector was expected to act (TEQSA, Gen AI strategies for Australian higher education: Emerging practice). 

Those plans revealed a sector still in flux: some providers had made substantial progress in redesigning assessment, while others were only beginning to grapple with the implications of AI. The new resource builds on that momentum, signalling that incremental adjustments won’t be enough.

Assessment design has to safeguard integrity, but it also has to reflect the skills and practices learners will need in the workplace. That dual purpose—assurance and preparation—is what makes assessment reform complex. It’s also why TEQSA has moved from principles to pathways, encouraging institutions to think beyond individual tasks and toward whole-of-program coherence.

The three pathways

TEQSA outlines three models that institutions are already experimenting with:

• Program-wide reform: redesigning assessment across an entire degree. This creates coherence, allows for progression in complexity, and aligns with graduate capabilities, but requires cross-faculty coordination and a cultural shift. In practice, program-wide approaches are often seen in professional fields such as medicine or law, where providers are mapping competency development over several years rather than relying on isolated high-stakes exams.
  Unit-level assurance: embedding at least one secure task in every subject. This provides immediate visibility for integrity but can lead to fragmentation if not connected to program outcomes.
• Hybrid models: combining program-level oversight with unit-level checks. This approach balances pragmatism and ambition, often acting as a transition pathway toward more systemic reform.

What matters is not the model itself but the strategic choice behind it. Each pathway comes with trade-offs in workload, governance, and learner experience. TEQSA’s message is clear: institutions need to be deliberate, aligning assessment approaches with both regulatory requirements and educational purpose.

Questions universities should be asking

TEQSA doesn’t provide prescriptive answers, but it does pose reflective questions that signal where institutions should focus:

• How do subject-level assessments connect to program outcomes?
• How can immediate integrity concerns be balanced with longer-term educational goals?
• What forms of support do staff and learners need to adapt to new assessment approaches?

These questions cut across context. Whether you’re redesigning a bachelor’s degree, a vocational qualification, or a school curriculum, the principles are the same: integrity, coherence, and authenticity.

Where Canvas and IgniteAI can help

Canvas already supports institutions taking program-wide, unit-level, or hybrid approaches. It provides the structures needed to align assessments with outcomes, track progress across learning journeys, and create reliable experiences for staff and learners.

IgniteAI, Instructure’s secure, in-context AI solution, builds on this foundation. It’s designed to help educators strengthen assessment design and delivery in ways that are authentic, sustainable, and aligned with regulatory priorities. That includes:

• Creating tasks that reflect professional practice and require original student effort.
• Providing timely, personalised feedback that supports learning while reducing workload.
• Enabling more inclusive experiences through features such as translation and adaptive prompts.
• Supporting assurance of learning across different models, rather than relying on detection of AI use.

Institutions have full control over how IgniteAI is deployed. Features are opt-in and configurable at the institutional level, giving leaders confidence that AI use aligns with their governance, compliance, and cultural priorities.

By extending the broader Canvas ecosystem, which already supports analytics, outcomes, and reporting, IgniteAI helps institutions design assessments that uphold integrity and prepare learners for the future.

Moving forward

TEQSA’s guidance is a signal to the sector that reform isn’t optional. Universities—and, by extension, schools and training providers—need to evolve their assessment strategies to ensure both compliance and relevance.

The opportunity lies in treating AI as a catalyst for redesign. Institutions that act now will be better positioned to protect the integrity of their qualifications, deliver high-quality learning experiences, and prepare learners for a world where AI is already part of study and work.