Daripada literasi kepada agensi: Bersedia untuk penglibatan kritikal dalam literasi media dan AI

Overview
At Bett Asia 2026, educational leaders addressed a fundamental question: how to prepare learners for an information environment reshaped by generative artificial intelligence (AI). Delivered by Lai Cheng Wong, Senior Communication and Publication Manager at SEAMEO RECSAM (the Southeast Asian Ministers of Education Organisation Regional Centre for Education in Science and Mathematics), the session titled "From Literacy to Agency: Preparing for Critical Engagement in Media and AI Literacy" outlined a crucial paradigm shift in educational priorities.

As synthetic media, deepfakes, and cloned audio proliferate, traditional media literacy habits—checking authors, web domains, or cross-referencing sources—remain necessary but are no longer sufficient. This overview synthesises the session's key insights, frameworks, and practical takeaways for educators, curriculum designers, and policy makers.

Key Takeaways at a Glance
From Source Verification to System Understanding: Verifying sources is still vital, but learners must now understand the underlying algorithmic systems that generate, edit, and recommend content.
Closing the Agency Divide: Equal access to AI tools does not produce equal learning. The central challenge in modern schooling is the agency divide—the gap between students who passively accept AI outputs and those who critically evaluate, adapt, or reject them.

Distributed Intelligence ("Effects With" vs "Effects Of" AI): Short-term efficiency gained by using AI ("effects with AI") must be distinguished from long-term intellectual growth ("effects of AI"), which strengthens human judgement over time.

The OECD PISA MAIL Framework: The OECD plans to introduce Media and AI Literacy (MAIL) in the 2029 PISA assessments across five competencies: Access, Analyse and Evaluate, Participate and Collaborate, Create, and crucially, Reflect and Act ethically.

The AAMMRR Discussion Lens: A practical classroom inquiry framework structured around Audience and Author (AA), Message and Meaning (MM), and Representation and Reasoning (RR).

The Evolving Challenge: Why Source Checking Is No Longer Enough
Digital misinformation has evolved rapidly. During the COVID-19 pandemic, educators focused on combating the "infodemic"—the spread of false news across social platforms. Today, generative AI presents more complex challenges through hyper-realistic deepfakes, automated text generation, and cloned media. Historically, media literacy taught students to evaluate content through simple checks: Who wrote this? What website is this from? Is the source reputable? In an AI-mediated environment, these habits remain essential, but are no longer sufficient. When synthetic content displays no obvious technical errors, evaluating credibility requires investigating the broader algorithmic systems, incentives, and business models driving its creation.
As highlighted during Bett Asia 2026, the smartest learner in the classroom is not the individual who retrieves the fastest answer from an AI model, but the learner who can discern, apply human judgement, and question AI-generated outputs.

Moving from the Access Divide to the Agency Divide
For years, digital education policy across Southeast Asia focused on overcoming the digital access divide by expanding internet connectivity, hardware provision, and technical skills training. While access gaps remain, generative AI has created a new inequality: the agency divide.

Consider two students using identical AI tools on the same network:
Student A inputs a prompt, receives an output, and uncritically adopts the response.
Student B inputs the same prompt, but asks: Is this accurate? What evidence supports this? What perspectives are missing? Should I adapt or reject this answer? Despite equal technical access, these two learners possess vastly different levels of agency.

Understanding Human Agency
Human agency in an AI-mediated world is the capacity to make informed, intentional choices rooted in personal and professional values. Exercising agency does not mean rejecting AI tools. Instead, it involves professional decision-making: knowing when technology adds value, adapting or rejecting automated suggestions, and explaining the pedagogical or ethical rationale behind those choices.

Distributed Intelligence: "Effects With AI" versus "Effects Of AI"
Preserving agency requires understanding distributed intelligence—a partnership where humans and AI tools think together. Educators must distinguish between immediate efficiency and long-term capability development:
Effects With AI: Operational efficiency gained while using a tool—such as an educator generating a lesson plan outline in ten minutes rather than an hour.

Effects Of AI: Lasting cognitive impact on the user over time. After months of working alongside AI, does the educator possess stronger critical questioning skills, better ability to spot weak evidence, and sharper independent judgement? Educational strategy must focus on the effects of AI, ensuring technology enhances human critical thinking rather than replacing it.

Evolving Teacher Roles in an AI-Mediated Environment
To foster student agency, educators must evolve into five key roles:

• Future Learning Architects: Designing experiences that integrate AI purposefully to stimulate deep thinking.
• Critical Mediators: Guiding students to interrogate algorithmic content and ask probing questions.
• Ethical Decision-Makers: Evaluating data privacy, fairness, and algorithmic bias in educational tools.
• Contextual Adapters: Tailoring content to suit diverse cultural, regional, and linguistic realities across Southeast Asia.
• Agency Builders: Supporting students to remain active, autonomous participants in their learning.

Teachers retain vital professional rights: deciding when AI adds value, choosing when to conduct tech-free lessons, adapting or rejecting AI recommendations, and explaining their pedagogical reasoning.

The OECD PISA MAIL Framework and the AAMMRR Lens
Looking ahead to 2029, the OECD intends to introduce Media and AI Literacy (MAIL) as a core assessment domain in PISA alongside reading and numeracy. The proposed framework outlines five competencies: Access, Analyse and Evaluate, Participate and Collaborate, Create, and Reflect and Act.

While all five dimensions are important, Reflect and Act ethically and responsibly form the central core. Students must learn to reflect on how AI influences their thinking and act responsibly to counter deepfakes, scams, and online harms within their communities.

Practical Application: The AAMMRR Inquiry Lens
To apply these principles in class, educators can use the AAMMRR framework for structured discussion:

Audience and Author (AA): Who or what created this content? Was AI involved in generation, editing, or curation? Who is the intended audience, and who benefits from it?

Message and Meaning (MM): Why am I seeing this message? Did an algorithm select it for me? What is its purpose or bias?
Representation and Reasoning (RR): How was this created? What data or sources were used? How does AI involvement affect the credibility of the message?

Cultivating a Culture of Critical Inquiry
Education in the generative AI era is fundamentally about human discernment, wisdom, and ethics. To build social resilience against digital harms, schools and policy makers must look beyond basic technology access and focus on developing human agency. By embedding Media and AI Literacy across curricula and encouraging structured inquiry through tools like AAMMRR, educational systems can ensure technology strengthens human critical thinking rather than replacing it.

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 di sini.