The integration of artificial intelligence into education is often described in terms of its potential: task automation, personalised learning, content generation… However, to define its true scope – beyond what these tools are capable of – it is becoming increasingly important to observe what teachers are already doing with them in their day-to-day practice.
Understanding how teachers use artificial intelligence – at what stages of the educational process it is used and for what purpose – allows us to contextualise the discussion about AI and its various applications in a more concrete way. What are teachers doing with AI?
A recent study, based on a survey of more than 500 practising teachers in the Dominican Republic, has enabled us to explore this question through the specific experiences of those who are incorporating these tools into their work. This article focuses on teachers’ actual use of artificial intelligence and on the conditions that determine its integration into teaching.
What are teachers doing with artificial intelligence?
The results show that the use of artificial intelligence is concentrated, above all, on lesson planning. According to the data, teachers use generative tools to prepare content, structure teaching units or create materials that are subsequently incorporated into lessons.
AI is thus linked to instructional design: drafting texts, devising explanations, generating examples or creating presentations are some of the most widespread uses. It is also beginning to be used to adapt activities to different levels or learning styles, one of the areas where these technologies show the greatest potential.
However, not all uses involve the same level of integration. In many cases, artificial intelligence is limited to speeding up tasks that were already part of a teacher’s role. In other, less frequent cases, it is beginning to be incorporated as part of the teaching process itself, influencing the sequence of activities or the way in which learning is tailored to students. The data indicate that the first type of use predominates, although there are early signs of more complex integration.
This pattern of adoption is no coincidence. Introducing a technology into the classroom requires time, familiarity and pedagogical judgement to decide when and how to use it. For this reason, artificial intelligence first finds its way into the areas where the teacher has the greatest control: lesson preparation. It is there that it is experimented with, adjusted and tested for its usefulness before being transferred, if appropriate, to the dynamics of the classroom.
What these figures show is not an abrupt transformation of teaching practice, but a more gradual change that begins at the design stage. Understanding this starting point allows us to better gauge the true scope of artificial intelligence in education. It is not enough simply to observe whether or not it is being used; we need to analyse at what stage of the educational process it intervenes and for what purpose. That is where the differences begin to emerge that will determine its impact.
Widespread, but uneven, use
The use of artificial intelligence tools by teachers is growing rapidly, but not at the same rate across all contexts. Are they being integrated in a similar way at different levels of education? Do they enable the development of new forms of teaching, or are they being incorporated into existing practices? The data point to an adoption process that is progressing at different rates.
Generally speaking, the level of perceived competence falls within the medium-to-high range, indicating that these tools are already part of many teachers’ repertoire. However, their use is not evenly distributed. The ability to use them and, above all, to integrate them into teaching varies significantly depending on the context.
Differences are evident, first and foremost, across educational levels. Early years teachers demonstrate the highest levels of competence in using generative tools to plan teaching, whilst in primary and higher education the results are lower or more inconsistent. This finding is, in a way, counterintuitive. One might expect the levels with greater subject-specific specialisation or greater exposure to technology to lead this process, but the opposite is true.
One possible explanation relates to the pedagogical conditions at each stage. At the early stages, teaching is usually organised more flexibly, with greater scope for experimentation and the adaptation of activities. In this context, tools capable of generating diverse resources (texts, images or audiovisual materials) fit in relatively easily. As pupils progress through the education system, the weight of the curriculum, standardised assessment and academic demands tend to reduce this scope for manoeuvre, which can limit the incorporation of new tools into everyday practice.
The differences are not limited to educational level. They also arise depending on the type of tasks that teachers are able to carry out using these technologies. The most widespread uses centre on text generation or the creation of presentations, whilst other applications, such as video editing or the use of more advanced visual tools, show lower levels of proficiency. This points to partial adoption: artificial intelligence is first integrated into tasks closer to everyday practices and progresses more slowly towards uses that require new skills.
This pattern suggests that the expansion of artificial intelligence in education does not follow a linear trajectory. Simply introducing a tool is not enough to ensure its use becomes widespread in the same way across all contexts. Factors such as the conditions of the education system, the demands of the curriculum or the teachers’ own capabilities all play a part.
In this scenario, it is more important to analyse how these tools are used than simply to measure their presence. The differences lie not so much in access as in the ability to integrate them into teaching practice. That is where their effects on teaching begin to take shape.
According to the data, teachers use generative tools to prepare content, structure teaching units or create materials that are then incorporated into lessons.
Training: the most significant factor
Whilst there are clear differences in the use of artificial intelligence in teaching, the data reveal one of the factors that best explains these differences: teacher training. Teachers who have taken part in training programmes in educational technology demonstrate higher levels of competence in virtually all the uses analysed. The difference is not marginal: it extends across a range of tasks, from planning activities to creating materials or adapting content.
This effect is particularly pronounced in secondary and higher education, where the gaps between teachers with and without prior training are more marked. At these levels, the use of generative tools requires, to a greater extent, the integration of subject-specific knowledge with pedagogical decisions. In this context, training emerges as a key element that enables these two aspects to be linked and allows technology to be incorporated into teaching practice more consistently.
The nature of use also changes. Among those who have received training, artificial intelligence is used more frequently to design activities tailored to different levels of learning or to organise more complex teaching sequences. In the absence of such training, its use tends to focus on more immediate tasks, such as generating texts or producing basic materials. The tool is present in both cases, but its role within the educational process differs.
This pattern points to a difference that goes beyond technical proficiency. Incorporating artificial intelligence into teaching involves making decisions about how to structure learning, which objectives to pursue, or how to adapt content to students. Training in educational technology not only fosters familiarity with the tools, but also provides criteria for using them in a pedagogically meaningful way.
The result is a shift in the nature of their use. As training increases, artificial intelligence ceases to be an ad hoc resource and becomes an integral part of instructional design. This change does not occur automatically or uniformly, but it introduces a significant difference in the way these technologies are integrated into teaching practice.
In this context, the expansion of artificial intelligence in education depends largely on the training opportunities available. Without them, its use tends to remain at more basic levels. With them, the scope for exploring more complex applications tailored to classroom needs is broadened.
Creativity as a key competence
Alongside training, the study identifies another factor with significant influence on the use of artificial intelligence in teaching: teachers’ creativity. The data show a consistent relationship between the level of creativity perceived by teachers and their ability to use generative tools in curriculum planning. This association holds true across all educational levels and reaches values that account for a significant proportion of the observed differences.
This link is no coincidence. Working with artificial intelligence often involves defining the desired outcome, fine-tuning the results and integrating them into a broader pedagogical approach. This process requires more than just technical skills. It involves imagining possible uses, adapting resources to specific contexts and making decisions about how to integrate them into learning. In this regard, creativity acts as a resource that broadens the scope of application.
Teachers with higher levels of creativity tend to use artificial intelligence in a more varied way and to incorporate it into tasks that demand a greater degree of elaboration. This can be seen, for example, in the design of personalised activities, in the production of multimodal materials, or in the combination of different formats within a single teaching sequence. In these cases, the tool ceases to be merely a content generator and becomes part of a broader process of pedagogical construction.
The difference is also evident in the type of tasks that are carried out with greater ease. Applications requiring more active intervention (such as image editing, the creation of audiovisual resources or the adaptation of complex content) are more closely linked to creativity than those related to text generation. As design demands increase, creativity takes on a more significant role.
This finding introduces an important nuance to our understanding of teachers’ digital competence. It is not enough simply to know how to use a tool or to be familiar with its functions. The ability to integrate it into teaching also depends on skills associated with design, experimentation and adaptation. In this sense, creativity emerges as a dimension that helps to give pedagogical meaning to the use of artificial intelligence.
What changes when artificial intelligence enters the classroom
As artificial intelligence is incorporated into teaching practice, the factors explaining its use also change. For years, the educational debate on technology has been dominated by access: the availability of devices, connectivity or platforms. The data from this study point in another direction. Differences in the use of generative tools are explained not so much by geographical context as by teachers’ capabilities.
The implications are clear. If the use of these technologies depends largely on teachers’ training, their uptake may follow uneven trajectories even in settings where resources are available. In the absence of training, artificial intelligence tends to be confined to more limited tasks. Where there is greater capacity building, the scope for integrating it into lesson design and exploring more complex applications widens.
This shift also affects the way in which educational innovation is understood. The introduction of new tools does not, in itself, guarantee changes in practice. What determines their scope is the way in which they are incorporated into existing processes and their capacity to transform them. Artificial intelligence does not operate on the fringes of teaching; it is embedded within it and takes on the forms permitted by those who use it.
In this context, observing how teachers are using these technologies offers a more accurate indication of their impact than any projection of their potential. What today appears to be a support tool for planning may, over time, become a more structural element of the educational process. The trajectory will depend less on the evolution of the tool than on the conditions under which it is integrated.
Understanding this point is key to anticipating what might change in teaching. Artificial intelligence is already part of a teacher’s work. Its impact will not be determined by its mere presence, but by the way in which it is used.
