How educators are actually integrating AI into their teaching

How educators are actually integrating AI into their teaching

Artificial intelligence is now a constant presence in higher education conversations.

It appears in policy documents, conference programs, assessment guidelines, and staff development initiatives. Yet amid the noise, a quieter and more important question often gets lost: what does meaningful integration of AI actually look like in teaching practice?

Across disciplines – from business and STEM to health, design, and education – educators are already integrating AI in ways that are thoughtful, pedagogically led, and deeply connected to learning goals. These integrations are rarely flashy. They don’t hinge on the latest tool or a perfectly engineered prompt. Instead, they reflect educators making deliberate design choices about how AI can support learning, extend practice opportunities, and strengthen student capability, while keeping human judgment at the centre.

What follows are some of the common patterns emerging from practice.

AI as a practice environment for developing skills

A common way educators are integrating AI is by using it to create safe, repeatable practice environments for developing complex skills. This approach is particularly evident in disciplines where learning involves interpersonal interaction, professional judgement, and confidence – skills that are difficult to practise at scale and often carry high stakes.

One example comes from Rachel Clark, a Social Care and Health educator from RMIT University’s College of Vocational Education, who works with students preparing for industry-based placements and work roles involving online service support, such as Lifeline and Beyond Blue Webchat. These roles require students to navigate complex, emotionally charged conversations while applying trauma-informed practice principles. Yet practising such scenarios live in class can present challenges. For some students, roleplaying sensitive situations with peers can trigger feelings of vulnerability  or draw on past personal experiences, making traditional practice environments difficult to manage.

To address this, Rachel introduced AI-supported roleplay as a practice environment designed around trauma-informed principles of safety, choice, and empowerment. By carefully prompting the AI, she created scenarios that reflected realistic client interactions at varying levels of complexity, allowing learning to be intentionally scaffolded. The AI could also be prompted to represent diversity across a range of client experiences, helping students develop professional awareness without placing emotional labour on classmates.

Crucially, the design gave students control over their learning. Students could practise independently, pause or stop an interaction at any time, and consult the educator when needed. This shifted practice from a performative classroom activity to a self-paced learning experience, where students determined their level of engagement and readiness. Rather than replacing in-class teaching, the AI-supported roleplay complemented observation, discussion, and educator-led guidance – creating a safer, more inclusive practice environment that supported both skill development and student wellbeing.

This pattern aligns with emerging research highlighting the value of AI-enabled roleplay as a tool for developing communication and counselling skills (Maurya, 2024). Studies have shown that AI-based roleplay environments can increase learner engagement, confidence, and reflective capacity when designed as formative, practice-oriented experiences rather than performance assessments (Freeman, 2025). By enabling repetition, variation, and immediate feedback, these environments support learning processes that are otherwise difficult to sustain within traditional teaching constraints.

Across contexts, the key insight remains consistent: AI works best as a practice partner when it is embedded within a broader pedagogical design. Educators shape the scenario, define what “good” looks like, and guide reflection. AI simply provides the space in which learning can be rehearsed, refined, and deepened.

AI as a coach, not an answer engine

One of the clearest patterns emerging in practice is the use of AI as a learning coach rather than a source of answers. Educators are deliberately designing AI interactions that support students through the learning process while preserving productive struggle, reflection, and judgment.

A strong example of this approach is the 3:00 AM Tutor developed by Thomas Bierley in RMIT University’s College of Business and Law. This Val (RMIT’s internal, secure AI Tool) persona  embedded directly into the learning environment supports students when human feedback is unavailable. Designed around established coaching frameworks, the tutor provides timely prompts, gentle hints, and encouragement, helping students stay in the “learning zone” rather than defaulting to solution-seeking. Crucially, it does not give answers. Instead, it guides learners to think, check their reasoning, and persist through difficulty.

In practice, this has enabled a shift away from static practice activities toward more dynamic, responsive learning experiences. Thomas’s colleague, Vineet Tawani, has also integrated the 3:00 AM Tutor in his finance course. The bot allows students to request additional practice on specific aspects of a calculation, receive tailored feedback, and iterate within clearly defined boundaries set by the educator. The result is not faster completion, but deeper engagement – students can maintain momentum, practise at their point of need, and refine understanding without waiting for instructor intervention.

This pattern aligns with emerging research showing that AI can be most effective when integrated as a formative, process-oriented support rather than a content generator. Studies in programming and language learning contexts have found that AI-assisted environments which scaffold thinking, provide feedback, and encourage iteration can improve learning outcomes while maintaining learner agency (Kuramitsu et al., 2023; Wang & Feng, 2023). Rather than replacing instructional intent, these designs extend it – enabling feedback-rich learning experiences that would otherwise be difficult to sustain at scale.

What is notable across these implementations is the pedagogical clarity underpinning them. Educators are not relinquishing expertise to AI; they are encoding their teaching intentions into the learning design. AI becomes a conduit for coaching, feedback, and persistence – supporting learning without short-circuiting it.

AI supporting educator design work behind the scenes

Not all meaningful AI integration is visible to students. Many educators are using AI behind the scenes to support their own teaching practice. This includes creating or refining learning materials, exploring alternative learning activities, or generating practice tasks that can then be carefully scaffolded into delivery.

This pattern is particularly evident in highly technical courses, where students can quickly feel overwhelmed if complex concepts are introduced without sufficient structure. In one such course, Khuong Nguyen, Lecturer for Electronic & Computer Systems Engineering  at RMIT University, explored how AI could support the design of scaffolded learning activities that would help students build confidence and capability over time. Rather than using AI to generate content, he used it as a creative partner, prompting for activity ideas and then refining, adapting, and positioning them using his own pedagogical and disciplinary expertise. The resulting activities were deliberately placed by Khuong within the course to support progressive skill development. 

What stands out in this example is the nature of the human–AI collaboration. AI contributed ideas and possibilities, but the educator made the decisions that mattered: what to include, how to scaffold it, and when to introduce it. Students experienced the outcome of this design work as more engaging, dynamic learning tasks that reduced cognitive overload and supported deeper understanding. Rather than feeling overwhelmed, they reported increased motivation and confidence as they progressed through the course.

In these cases, AI functions as a design assistant rather than a content authority. Educators draw on their disciplinary expertise and knowledge of their students to curate, adapt, and contextualise what AI suggests. The pedagogical decisions remain firmly human. AI simply helps educators work more efficiently and creatively within the constraints of time, workload, and curriculum structure.

A shift led by educators

Taken together, these practices tell a clear story. Meaningful AI integration is not about replacing teaching or chasing technological novelty. It is about educators exercising professional judgment, making intentional design choices, and adapting their practice to reflect the realities of an AI-enabled world.

For some, this integration is already visible in the classroom. For others, it may be happening quietly in planning documents, assessment redesigns, or small shifts in learning activities. Both matter.

Perhaps the most useful question for educators is no longer “Should I be using AI?” but rather: Where is AI already intersecting with my teaching practice and how can I shape that intersection more intentionally to support learning?

02 March 2026

More GAILE blogs

aboriginal flag float-starttorres strait flag float-start

Acknowledgement of Country

RMIT University acknowledges the people of the Woi wurrung and Boon wurrung language groups of the eastern Kulin Nation on whose unceded lands we conduct the business of the University. RMIT University respectfully acknowledges their Ancestors and Elders, past and present. RMIT also acknowledges the Traditional Custodians and their Ancestors of the lands and waters across Australia where we conduct our business - Artwork 'Sentient' by Hollie Johnson, Gunaikurnai and Monero Ngarigo.

Learn more about our commitment to Indigenous cultures