UNESCO proposes five units for students aged 13 to 15 to question AI’s answers and decide for themselves
“AI answers. You decide” is a companion guide to UNESCO’s media and information literacy curriculum, with classroom discussions and activities that can be used separately.
In 2026 UNESCO published “AI answers. You decide: A companion guide to the UNESCO media and information literacy curriculum”, a guide written by Anne Leppäjärvi, a Finnish educator in journalism and media and information literacy, and coordinated by Adeline Hulin and Olivier Van Bockstael of the organisation’s staff. The guide itself warns that the ideas expressed are those of its authors and do not necessarily commit UNESCO. It is in English and open access, under a CC-BY-SA 3.0 IGO licence.
It is aimed at teachers, not students, and requires no technical knowledge. It complements the curriculum “Media and Information Literate Citizens: Think Critically, Click Wisely!” (second edition, 2021) and UNESCO’s AI competency framework for teachers, whose media and information dimension, the guide says, it addresses. The activities are designed mainly for young people aged 13 to 15, with suggestions for adapting them to younger or older ages. Each section can be taught as a stand-alone lesson, so there is no need to follow the guide from start to finish.
Its guiding thread is the “Orientation Map”, with three layers: the individual, the AI systems, and the people and institutions that build, fund and regulate them. Alongside it, it proposes five questions that recur in every unit: who made it, what evidence supports it, what system shaped it, who benefits from you believing or sharing it, and what should I verify before acting. The first unit distinguishes four things AI does (predict, classify, make decisions about a person and generate content) and deals with the data people give up when using a free service. The second presents AI as both a producer and a curator of content, and proposes comparing four types of source according to their degree of editorial responsibility.
The third unit is the closest to everyday use of chatbots. It explains that a language model predicts plausible text from patterns, that fluency is not the same as accuracy, and that a synthesised answer can hide the sources and disagreements that a search engine does show. It offers the chain “trust, verify or question” depending on what is at stake, three self-check questions (“am I learning or just getting a result?”) and a four-step research loop. Its working rule is to treat an AI answer as a lead, not a source: the guide says, verbatim, “treat an AI-generated answer as a lead rather than a source”. It also devotes a section to those who seek companionship from a chatbot and, in that case, recommends a private, supportive conversation.
Units four and five turn to power, rights and the future. They address inequality that goes beyond access (language, computing power, knowledge represented in the data, cost), the accountability of algorithms, the energy cost of AI with figures from UNESCO research from 2025 that, according to the guide itself, will change, and three development scenarios: a few companies decide, public AI, and many small developers. Mariya Gabriel’s foreword sums up the approach in one sentence: “AI should not simply happen to us”. One of its arguments is that not using AI is a legitimate decision, as long as it does not mean giving up on understanding it.
It has limits worth knowing about. It is a long text, in English, written from an international perspective, and the examples and legal frameworks it cites do not always match the Spanish context. Several activities ask students to use an AI tool, and the guide reminds readers to check the school’s policy and each service’s minimum age beforehand; it cites UNESCO’s 2023 recommendation to set 13 as the age for independent use. The Diario has already looked at the question of age in a report for UNESCO on the appropriate age. The guide also notes that most of the activities can be done with the teacher operating the tool on a shared screen.
For a teacher, the most useful approach is to start with the “Discussion Guide” and “Try This” sections of the unit that best fits their subject. Activities such as “Sort My Day” (classifying the day’s uses of AI) or “Ask the Three Questions” (applying three questions to a chatbot’s answer and checking one of its sources) fit into a single session.
UNESCO proposes five units for students aged 13 to 15 to question AI’s answers and decide for themselves
“AI answers. You decide” is a companion guide to UNESCO’s media and information literacy curriculum, with classroom discussions and activities that can be used separately.
In 2026 UNESCO published “AI answers. You decide: A companion guide to the UNESCO media and information literacy curriculum”, a guide written by Anne Leppäjärvi, a Finnish educator in journalism and media and information literacy, and coordinated by Adeline Hulin and Olivier Van Bockstael of the organisation’s staff. The guide itself warns that the ideas expressed are those of its authors and do not necessarily commit UNESCO. It is in English and open access, under a CC-BY-SA 3.0 IGO licence.
It is aimed at teachers, not students, and requires no technical knowledge. It complements the curriculum “Media and Information Literate Citizens: Think Critically, Click Wisely!” (second edition, 2021) and UNESCO’s AI competency framework for teachers, whose media and information dimension, the guide says, it addresses. The activities are designed mainly for young people aged 13 to 15, with suggestions for adapting them to younger or older ages. Each section can be taught as a stand-alone lesson, so there is no need to follow the guide from start to finish.
Its guiding thread is the “Orientation Map”, with three layers: the individual, the AI systems, and the people and institutions that build, fund and regulate them. Alongside it, it proposes five questions that recur in every unit: who made it, what evidence supports it, what system shaped it, who benefits from you believing or sharing it, and what should I verify before acting. The first unit distinguishes four things AI does (predict, classify, make decisions about a person and generate content) and deals with the data people give up when using a free service. The second presents AI as both a producer and a curator of content, and proposes comparing four types of source according to their degree of editorial responsibility.
The third unit is the closest to everyday use of chatbots. It explains that a language model predicts plausible text from patterns, that fluency is not the same as accuracy, and that a synthesised answer can hide the sources and disagreements that a search engine does show. It offers the chain “trust, verify or question” depending on what is at stake, three self-check questions (“am I learning or just getting a result?”) and a four-step research loop. Its working rule is to treat an AI answer as a lead, not a source: the guide says, verbatim, “treat an AI-generated answer as a lead rather than a source”. It also devotes a section to those who seek companionship from a chatbot and, in that case, recommends a private, supportive conversation.
Units four and five turn to power, rights and the future. They address inequality that goes beyond access (language, computing power, knowledge represented in the data, cost), the accountability of algorithms, the energy cost of AI with figures from UNESCO research from 2025 that, according to the guide itself, will change, and three development scenarios: a few companies decide, public AI, and many small developers. Mariya Gabriel’s foreword sums up the approach in one sentence: “AI should not simply happen to us”. One of its arguments is that not using AI is a legitimate decision, as long as it does not mean giving up on understanding it.
It has limits worth knowing about. It is a long text, in English, written from an international perspective, and the examples and legal frameworks it cites do not always match the Spanish context. Several activities ask students to use an AI tool, and the guide reminds readers to check the school’s policy and each service’s minimum age beforehand; it cites UNESCO’s 2023 recommendation to set 13 as the age for independent use. The Diario has already looked at the question of age in a report for UNESCO on the appropriate age. The guide also notes that most of the activities can be done with the teacher operating the tool on a shared screen.
For a teacher, the most useful approach is to start with the “Discussion Guide” and “Try This” sections of the unit that best fits their subject. Activities such as “Sort My Day” (classifying the day’s uses of AI) or “Ask the Three Questions” (applying three questions to a chatbot’s answer and checking one of its sources) fit into a single session.
