Students are already using AI to search, write, translate, summarize, generate images and solve problems. But simply giving learners access to AI does not create AI literacy.
The real question for schools is more fundamental: What should a child understand about AI at age five, at age ten, at age fifteen, and by the time they leave school?
A meaningful AI-literacy program cannot be a single course taught once. It needs to develop progressively, just like mathematics, science or language.
That means moving from awareness and simple concepts in the early years toward critical evaluation, responsible use, system design and governance in secondary school.
The UAE is already moving in this direction. Its National Education Charter includes mastery of artificial intelligence among the future skills expected of learners, while the Ministry of Education has established a National AI Literacy Curriculum Framework. Abu Dhabi’s ADEK has gone further by implementing a progressive KG-to-Grade-12 AI Literacy initiative across more than 170 private schools, with students moving from early critical-thinking experiences toward evaluating AI outputs, building AI-powered solutions and understanding responsible governance.
So what should that progression actually look like?
KG: discovering intelligent machines
For young children, AI literacy should not begin with algorithms or chatbots.
It should begin with simple questions: What is a machine? What makes something “smart”? Who gives machines instructions? Are people and machines the same?
At this age, children can explore AI through stories, sorting games, sequencing activities, role-play and simple examples from everyday life.
The objective is not technical knowledge. It is to establish a few essential ideas: humans create technology, machines follow instructions, machines can recognise patterns, AI can help people and machines do not think or feel in the same way humans do.
This is also the right stage to introduce simple digital-safety habits.
Grades 1-2: patterns, examples and mistakes
The next step is helping children understand that AI can recognise patterns.
Students might classify objects by colour or shape, explore how image recognition works, or see how a system learns from many examples.
By Grade 2, another important concept can emerge: AI can make mistakes. Students can begin exploring how unclear instructions or poor examples can lead to incorrect results, and why humans need to check what machines do.
This is an early form of critical AI literacy.
Grades 3-4: data, rules and decisions
Students can now move beyond recognizing AI toward understanding some of its mechanisms. They can explore:
- data and examples;
- simple algorithms;
- “if-then” rules;
- predictions from past patterns;
- good and poor-quality data;
- privacy and personal information.
By Grade 4, the question becomes: How does AI use rules, data and patterns to make decisions and predictions?
The important shift is from seeing AI to beginning to understand why it behaves the way it does.
Grades 5-6: how AI learns
In upper primary, students can start exploring machine learning more directly.
They can investigate classification, training data, supervised and unsupervised learning, simple neural-network concepts and measures of accuracy. They can also begin asking more critical questions: What happens if the training data is incomplete? Can AI be unfair? How do we know whether an AI system is accurate?
By Grade 6, a strong central question is: How does AI learn from data?
This is also an ideal stage for hands-on projects in which students design a simple AI solution around a real-world or community problem.
Grades 7-8: evaluating AI
By lower secondary school, students are capable of moving from understanding AI to evaluating AI.
This is where concepts such as bias, deepfakes, misinformation, fairness, privacy and generative AI become increasingly important. Students should learn that an AI system can be technically accurate while still raising ethical questions.
They can start comparing outputs, examining data quality and asking: Is this result reliable? Is it fair? What evidence supports it? Who could be affected by the decision?
At this level, AI literacy begins to intersect strongly with media literacy, digital citizenship and critical thinking.
Grade 9: LLMs, prompting and trust
By Grade 9, students are already likely to be using large language models.
The educational objective should therefore not simply be to teach them how to prompt. It should be to teach them how to challenge an AI response.
At this stage, students can experiment with prompt design, context, role and perspective, hallucinations, source verification, bias, academic integrity and responsible AI use.
A key question becomes: How do we use and challenge LLMs responsibly?
Students should understand that a fluent answer is not necessarily a correct answer, and that confidence is not evidence.
Grades 10-11: comparing, designing and governing AI
At this stage, students can engage with more sophisticated systems. They can explore neural networks, deep learning, model performance, explainability, accessibility, cybersecurity, high-stakes decision-making and human-AI collaboration.
They should also start considering how different AI systems compare.
Which model is best suited to which task? What trade-offs exist between performance, fairness and explainability? Who is accountable when an AI system causes harm?
By Grade 11, the central question moves toward: How do we make AI trustworthy, inclusive and accountable?
This prepares students not only to use AI, but also to understand the responsibilities involved in designing and deploying it.
Grade 12: AI as a societal system
By the end of school, AI literacy should go beyond tools and models.
Students should be able to examine AI through ethical, societal, economic and governance lenses. They can explore:
- AI regulation and compliance;
- data governance;
- auditing;
- environmental impact;
- future careers;
- public policy;
- AI in healthcare, transport and government;
- humanitarian and global challenges.
A strong Grade 12 capstone should ask learners to design, evaluate and present an AI-enabled solution while considering both technical performance and societal impact.
The final question becomes: How do we design, govern and deploy AI responsibly in society?
A coherent K-12 progression
Seen as a whole, the progression can be summarized simply: Discover AI → Understand AI → Use AI → Question AI → Evaluate AI → Design with AI → Govern AI responsibly
This is close to the direction now being adopted in the UAE. ADEK, for example, explicitly describes a progression from early awareness and critical thinking to AI solution building and, by Grade 12, systems and solutions design.
The challenge for schools, however, is turning frameworks into something teachers can actually use.
From framework to classroom: the Mexty editable curriculum
This is why Mexty has developed a complete editable AI Literacy curriculum templates from KG to Grade 12, designed to help schools move from high-level objectives to practical implementation.
The purpose is not to impose one rigid curriculum. It is quite the opposite. Schools need flexibility because they operate under different national curricula, teaching models, timetables and student profiles.
The Mexty curriculum is therefore designed as an editable starting point.
Schools can adapt the structure, modify lessons, add their own policies and examples, change assessments, integrate local context, and align content with their own curriculum requirements.
For each grade, the template can combine: learning paths, courses, interactive activities, AI experiments (AI Literacy Lab), assessments and projects.
Teachers can retain control over the content while adapting or extending it.
This also makes it possible to integrate practical AI-literacy experiences into the curriculum where students can compare answers from different AI models, explore how context changes an output, investigate hallucinations, examine bias, or compare general model knowledge with responses grounded in trusted sources.
The goal is to move from simply teaching students about AI toward allowing them to experience, question and evaluate AI in a structured and secure environment.
Mexty’s role is therefore to provide schools with a practical framework and structured foundation and the secure learning infrastructure to adapt, deliver and assess AI Literacy curriculum, while teachers and school leaders retain pedagogical control.
AI literacy is a journey, not a single subject
Perhaps the biggest lesson is that AI literacy should not begin when students first encounter ChatGPT and it should not end when they learn how to write a good prompt.
It is a progressive capability to be developed over many years.
A five-year-old needs to understand that machines follow human instructions.
A ten-year-old can begin understanding how data and patterns influence AI.
A fourteen-year-old should be able to question an AI-generated answer.
And an eighteen-year-old should be capable of thinking critically about how AI systems are designed, evaluated and governed.
That is the progression schools now need to build and with national frameworks becoming increasingly ambitious.
It is turning that ambition into a coherent learning journey from KG to Grade 12.
