In recent years, teaching has become denser work. More administrative tasks, more production of materials, more adaptation to different classroom rhythms. In that context, artificial intelligence began to be used as a practical tool before it became a pedagogical issue. Many teachers adopted it to save time on specific tasks: drafting activities, reorganizing content, or preparing assessments. It was not a curricular decision but an everyday response to an increasing workload.
That origin explains part of the current debate. AI first spread as individual support, not as a shared educational change. Teachers experiment with it, but this experimentation rarely becomes part of a collective reflection on what truly changes in learning. The result is an ambivalent situation: interest in technological possibilities accompanied by doubts about its role in the classroom and the limits of its use.
It is precisely in this context that the report Recommendations for Teacher Training in Artificial Intelligence in Basic Education is situated. The document begins with a diagnosis: technology has arrived before the training frameworks necessary to integrate it with pedagogical meaning. Its intention is to organize a discussion that has so far developed in fragmented ways, centered on individual use rather than professional development.
Teaching has always meant deciding which processes deserve time and effort because they are part of learning. When a tool allows intellectual tasks to be automated—such as summarizing, structuring texts, or generating answers—parts of the process that previously supported learning may disappear. What happens to teaching then? How do we redefine the work of teaching when part of the cognitive effort can be delegated so easily?
The report starts from that shift. Instead of focusing on how to introduce AI into schools, it centers attention on the teacher training required to coexist with it. The focus moves away from tool use and toward professional capacity to set boundaries, prioritize processes, and decide when technology enhances learning and when it oversimplifies it.
We Need More Judgment, Not More Tools
Teacher technology training has followed the same pattern for years. Each new tool came with courses focused on use: how it works, what it allows, how to integrate it into a specific activity. Artificial intelligence seemed destined to follow the same path. However, the report proposes a change: the issue is not that teachers need to learn more tools, but that they need stronger criteria to decide when to use them.
The proposal is organized around four dimensions:
- Teaching about AI (understanding what it is, how it works, and its limits).
- Teaching with AI (using it to support pedagogical practice).
- Understanding its social, ethical, and cultural implications.
- Integrating AI into teachers’ own professional development.
This distinction is crucial. Knowing how to use a tool is not the same as knowing how to integrate it pedagogically. Technological literacy can be learned relatively quickly; pedagogical judgment requires time, experience, and shared reflection. AI training is not a separate field, but an extension of existing digital teaching competencies.
AI thus ceases to be an added innovation and becomes a transversal technology embedded in planning, assessment, material production, and information management. Training stops being an isolated course and becomes part of a continuous professional process.
AI does not require a new type of teacher. It requires a teacher with greater capacity to decide what role technology should play in learning and which processes remain irreplaceable.
Training as Educational Policy
The debate often focuses on the individual teacher’s digital competence. The document introduces another scale: AI training depends on institutional decisions that create conditions for continuity and coherence.
Teachers do not start from the same point. Differences in digital experience, infrastructure, time, and pedagogical priorities matter. Homogeneous training designs often fail because they ignore this diversity. Training must begin by understanding real school contexts.
This approach shifts responsibility from individuals to organizations. AI integration cannot rely solely on isolated experimentation. It requires peer collaboration, communities of practice, pedagogical support, and institutional time to test and adjust practices.
Training also needs follow-up—not as control, but as part of professional learning. Observing how practices evolve allows continuous adjustment.
AI becomes a broader case study: innovation fails when it is treated as an individual task. It succeeds when it becomes organizational and policy-driven.
AI, Curriculum, and Equity
The report insists on integrating AI into the existing curriculum rather than adding it as another subject. AI connects with understanding data, automated decision-making, and technological implications.
In Brazil, this connects with the BNCC Computação framework, structured around computational thinking, digital world, and digital culture. Students need conceptual foundations before using intelligent systems.
Access alone is not enough. Schools differ in connectivity and infrastructure. Linking AI exclusively to tool use risks widening inequalities. The report therefore emphasizes “unplugged AI”: activities that teach concepts like pattern recognition and classification without digital devices.
Understanding systems, their limits, and the human decisions behind them may have more educational impact than intensive tool use. Integrated into the curriculum, AI becomes part of a broader conversation about learning in a digital world shaped by data and automation.
What Changes in the Conversation About AI and Education
AI has entered schools without waiting for clear guidance. The debate is shifting: less about possibilities, more about decisions.
The central question is not what AI can do, but who decides when to use it, for what purpose, and with what limits. In that context, teacher training becomes the space where professional judgment is built in an environment of constant change and no stable solutions.
