Assuring quality learning in a gen AI-integrated future: The role of adaptive capabilities. *TEQSA*, June 2026

Synthesis: Lodge et al. (2026), a TEQSA-commissioned report, proposes that assuring quality learning in a gen-AI-integrated future depends on cultivating four adaptive capabilities in graduates: digital literacy, distributed cognition, hybrid metacognition, and life-long learning — all supported by agency and AI Regulation in Education at the center. The report argues these durable, human-centered capabilities (built on deep disciplinary knowledge) let graduates maintain agency and use gen AI as a partner rather than becoming dependent on it. It offers five propositions of practice: make adaptive capabilities core graduate attributes, build institutional infrastructure for learning process evidence, design environments that promote them, shift toward process-focused assessment, and foster cross-institutional collaboration. This is a significant quality-assurance and policy framework for AI-mediated Australian higher education.

Key Findings

1. Four adaptive capabilities define a human-centered foundation for gen-AI-integrated learning. Digital literacy, distributed cognition, hybrid metacognition, and life-long learning — supported by agency and regulation — enable graduates to navigate complex, novel and gen-AI-integrated environments by deploying, adapting and transferring skills effectively and ethically.
2. Agency and regulation sit at the center. The human capacity to plan, monitor, evaluate and adjust learning and decision-making (self-regulated and co-regulated learning) underpins all four capabilities, letting learners maintain autonomy and manage Cognitive Offloading in complex contexts.
3. Five propositions re-orient practice. Graduate attributes, learning-process-evidence infrastructure, environment design, process-focused assessment, and cross-institutional collaboration collectively translate the Threshold Standards into actionable mechanisms.
4. The shift is from products to processes. Assuring learning quality requires evidencing how students learn (traces, regulatory activity, gen-AI interactions) — not just final products and outcomes — because gen AI can generate many traditional assessment products.

The four adaptive capabilities

• Digital literacy — using digital tools (including gen AI) effectively, ethically and safely; extends the Australian Digital Capability Framework (DigComp-based) with a critical understanding of gen AI's principles, methods, limitations, ethics, societal impact and power structures.
• Distributed cognition — how cognitive processes and tasks are shared across people, tools, artifacts and gen AI systems: working in human–gen AI teams, using external representations and agents to extend thinking, understanding knowledge flow across distributed systems, coordinating roles and resources.
• Hybrid metacognition — the regulation of thinking and learning within any cognitive system, including human–gen AI networks: planning, monitoring, evaluating and adjusting cognitive processes, whether in the individual or distributed. Ensures learners maintain agency, intentionality and ethical awareness (evaluative judgment, deciding when to rely on or override gen AI, co-regulation, ethical reasoning).
• Life-long learning — sustaining motivation, capability and adaptability to learn continuously in evolving, uncertain, gen-AI-mediated contexts: adapting to evolving technologies, identifying knowledge gaps, transferring learning, developing a learner identity grounded in agency.

Five propositions of practice

1. Establish adaptive capabilities as core graduate attributes. Embedding the four capabilities operationalizes Threshold Standards graduate-attribute requirements across disciplines, aligned with the National Skills Taxonomy and Australian Digital Capability Framework. Professional learning for educators focuses on Scaffolding students' adaptive capabilities within human–gen AI networks.
2. Build institutional infrastructure for learning process evidence. Invest in systematic capacity to collect and analyze data on learning processes — traces students generate as they plan, monitor, seek feedback, revise, and interact with gen AI (time-stamped edits, resource-use patterns, Problem Solving sequences, metacognitive prompts, structured self-assessments) — prioritizing student Learner Agency, privacy and transparency (e.g., open learner models).
3. Design learning environments that promote adaptive capabilities through evidence-informed pedagogical practices, including desirable difficulties and the value of productive struggle.
4. Transform pedagogical practice toward process-focused assessment. Rebalance assessment so learning-process documentation complements (not replaces) product evaluation. Learning analytics make processes visible; research (e.g., Raković et al. 2023) shows process features can explain more variance in performance than product features alone. Requires training educators and fading scaffolding as students develop self-regulated learning.
5. Foster collaborative innovation across institutions. Cross-institutional collaboration shares resources, research findings and approaches, building sector-wide capacity while prioritizing human agency and educational effectiveness over technological capability.

What this means for practice

• Instructors. Design learning environments around desirable difficulties and productive struggle, and fade Scaffolding as students develop Self-Regulated Learning, because the report's mechanisms are meant to be deliberately cultivated rather than assumed.
• Instructors. Rebalance assessment so documentation of learning processes complements — not replaces — evaluation of final products, since gen AI can generate the products assessment has traditionally relied on; the report treats learning processes as the durable evidence gen AI cannot simulate, reinforcing Assessment Validity and Authentic Assessment.
• Administrators. Establish the four adaptive capabilities — digital literacy, Distributed Cognition, hybrid metacognition, and life-long learning, with agency and regulation at the center — as core graduate attributes tied to the Threshold Standards' graduate-attribute requirements.
• Administrators. Invest in institutional infrastructure for learning-process evidence: the traces students generate as they plan, monitor, seek feedback, revise, and interact with gen AI, with student Learner Agency, privacy, and transparency (e.g., open learner models) designed in from the start.
• Administrators. Fund cross-institutional collaboration and educator professional learning that prioritize human agency and educational effectiveness over technological capability.

Limitations

• This is a commissioned policy and quality-assurance report, not an empirical study: it offers no participant sample, intervention, or outcome measure, so it can argue that adaptive capabilities matter but cannot show that cultivating them improves learning.
• Its central evidence for process-focused assessment is cited from other work — Raković et al. 2023 on multi-text writing, where process features explained more variance than product features — rather than generated by the report itself.
• The framework is grounded in the Australian Higher Education Standards Framework and Australian national frameworks, and the advice is explicitly non-prescriptive, leaving its fit to other jurisdictions and standards regimes untested.
• The five propositions are proposals: no institution-level implementation or evaluation data are offered, so their feasibility, cost, and effect at scale are unaddressed.

Citation

Lodge, J. M., de Barba, P., Ainscough, L., Brazil, J. R., Broadbent, J., Ebbert, D., Frankland, S., Gabriel, F., Gašević, D., Hennicke, T., Lim, L.-A., Male, S. A., Mirriahi, N., Oliveira, E. A., Pacitti, H., Raković, M., Russell, J., Taylor-Griffiths, D., & Yang, S. (2026). Assuring quality learning in a gen AI-integrated future: The role of adaptive capabilities. TEQSA, June 2026.