Driving real impact with AI: Teacher readiness, effective integration and student out

Summary
Artificial Intelligence (AI) integration in education has reached a pivotal inflection point across the Asia-Pacific region. While survey data indicates near-universal adoption amongst educators, school leaders face a fundamental challenge: converting raw technological usage into genuine, measurable learning gains. During a panel discussion at Bett Asia 2026, entitled “Driving real impact with AI: Teacher readiness, effective integration and student outcomes”, industry analysts, institutional leaders, and pedagogical experts laid out a strategic roadmap for meaningful AI integration.

The panel emphasised that AI must expand teacher capacity rather than replace professional judgement. To achieve sustainable outcomes, educational institutions must transition from tool-centric adoption to pedagogy-first frameworks, supported by ongoing, holistic professional development and risk-proportionate governance models.

1. The Adoption Paradox: Moving Beyond the "Fire Hose"
Recent market research presented by Futuresource Consulting reveals that 90 per cent of surveyed teachers across the region report using AI in their educational practice. However, high adoption rates mask a growing operational challenge described by Simon Hornbrook as "trying to drink from a fire hose". The rapid proliferation of disparate AI software creates friction regarding data privacy, student safeguarding, and evaluation.

To overcome this fragmentation, St Joseph's Institution International adopted an enterprise strategy. Rather than attempting to vet and deploy dozens of single-purpose utilities, the institution secured pro-level enterprise AI access across its entire community, including teachers, administrative staff, and students. This centralised approach established a secure environment where users developed foundational prompting skills and built customised tools without exposing sensitive institutional data.

2. Pedagogy First: AI as a Capacity Expander
A unifying theme across the session was the mandate to prioritise pedagogy over technology. AI should function as a high-speed computing assistant that automates repetitive tasks—such as lesson material adaptation, research synthesis, and administrative workflows—thereby freeing educators to focus on human connection and active instruction.

Key pedagogical principles highlighted by the panel include:

Task Performance vs Learning Gains: Jennifer Bantigue-Angeles highlighted that a polished AI-generated output does not guarantee deep learning. Generative tools can improve task execution without producing cognitive development unless guided by sound instructional design.

Making Thinking Visible: Dr Jerilyn Bacroya-Magbuo shared the "Human Library Project" at FAITH Colleges. Communication students conducted qualitative interviews with marginalised individuals—including street dwellers and ex-convicts—and subsequently developed interactive AI avatars. This exercise used AI to amplify empathy and make student thinking visible, demonstrating how technology can enhance human-centred learning.

Preserving Student Agency: Technology must not perform the critical thinking on behalf of the learner. AI should stimulate enquiry, challenge assumptions, and prompt students to ask higher-quality questions.

3. Redefining Teacher Readiness and Professional Development
One-off training sessions at the start of an academic year are inadequate for a technological landscape evolving daily. Research underscores a direct correlation between the frequency of professional development (PD) and teacher confidence, which in turn drives positive student outcomes. Panellists outlined several core components of effective AI PD:

Structured Four-Stage Scaling Framework: Jennifer Bantigue-Angeles recommended a disciplined model for scaling AI initiatives across campuses:

Pilot: Test tools with a small group of educators to gather empirical evidence.

Proof of Value: Verify that the tool genuinely enhances target learning outcomes.

Playbook Creation: Establish clear institutional guidelines defining non-negotiable standards (data safety, ethics) alongside flexible negotiables.

Contextualisation: Empower individual teachers to adapt strategies to their specific classroom needs.

Holistic Governance and Leadership Inclusion: Dr Jerilyn Bacroya-Magbuo noted that professional development is frequently offered to teachers or students while institutional leaders are omitted. School leaders require comprehensive training to draft robust policies and manage infrastructure effectively.

Data-Informed Customisation: Training must reflect internal readiness audits rather than generic, vendor-supplied packages. Tailoring PD to a school's infrastructure and socio-economic context ensures relevance and staff buy-in.

Critical Verification Skills: Educators and students must be trained to critically evaluate AI outputs for hallucinations, factual inaccuracies, algorithmic bias, and privacy compliance.

Eliminating "AI Tennis": Dr Magbuo cautioned against non-meaningful implementation cycles where teachers use AI to generate assignments, students use AI to construct responses, and teachers use AI to grade the submission—a process that eliminates genuine human engagement.

4. Risk-Proportionate AI Governance
To manage risk without stifling innovation, Simon Hornbrook outlined a governance framework calibrated to student impact:

Low-Risk Use Cases (Minimal Scrutiny): Internal brainstorming, administrative drafting, and preliminary idea generation.

Medium-Risk Use Cases (Moderate Scrutiny): Creation of student-facing learning materials and instructional resources, requiring rigorous verification of accuracy.

High-Risk Use Cases (Maximum Scrutiny): Tasks involving student personal data, well-being assessments, or formal academic reporting.

This continuum allows schools to maintain stringent data privacy and safeguarding standards while providing staff with clear boundaries for creative experimentation.

Audio and written summaries powered by Gemini Notebook, a free generative AI tool grounded in your provided sources. Try Gemini Notebook for free. You can listen to the audio overview of the session here.