Technology and education: 30 teachers working with Artificial Intelligence
Bianca Silva
By: Bianca Silva
Since the advent of ChatGPT, the debate over the use of Artificial Intelligence (AI) has become increasingly constant in the fields of technology, advertising, film, and also education. The question that stands out is: to what extent is access to Artificial Intelligence present in education when we consider the context of Brazilian public schools?
It was with this question in mind that we developed the MACMA 3.5 project, whose proposal was to bring together the knowledge of public-school teachers and the use of generative Artificial Intelligence to create 8,568 materials featuring active learning methodologies within three months, distributed free of charge on the AprendiZAP Platform.
The name MACMA 3.5 may sound intimidating at first, but it is the nickname we gave to “More activities with active methodologies.” If you’re wondering where the 3.5 comes from, I’ll tell you: this number corresponds to the version of ChatGPT we used to create the materials. So, now that we know the project’s name and what it means, let’s understand where the idea came from.
1. The problem in teachers’ lives: how can we make classes more engaging?
In interviews with teachers who use the AprendiZAP Platform, we found that, alongside a lack of time, there was an interest in offering stimulating and challenging activities to their classes.
Based on this discovery, our Content team conducted an investigation process in mid-August 2022, examining educational materials and platforms that offered guidance to teachers, providing guidelines on how to develop topics with their classes. Combining the interviews with this research, we defined a pilot project for creating these materials, which were to include:
active methodologies: to promote student agency and make classes more engaging;
a skill from the National Common Curricular Base (BNCC): so that teachers could identify which skill would be developed through the activity;
step-by-step instructions: a guide on how to implement the methodology, with tips for implementation.
Once the pilot was defined, we invited two specialists in Active Methodologies and Education—the educator and PhD Alice Ferreira and designer Camila Martinelli—to map which methodologies could be applied, taking into account difficulties accessing the internet and digital equipment. As a result, in October 2022 we mapped 12 active methodologies and created templates for each proposal, with the goal of providing teachers with guidance on how to develop the activity, as well as a teaching resource to use with their classes.
Template for the 3Q’s activity, used by students for self-assessment of their learning and/or evaluation of the class.
The templates were created for teachers to download and print, taking paper conservation into account (hence the dotted line in the middle, indicating where to cut) or to reproduce on the classroom board. Source: AprendiZAP.
To test the format of the materials, we selected 80 Portuguese Language topics from the AprendiZAP Platform and created the first teacher-guidance materials. During this period, between November 2022 and January 2023, we worked with specialist teachers responsible for analyzing the topic, the BNCC skill, and the methodology—among the 12 mapped—that best fit the content in question.
We prepared the materials, asked teachers to test them, gathered feedback, made adjustments, and made the 80 pieces of content available on our platform at the beginning of February 2023. Through this process, our content team was able to learn more quickly, understand which methodologies and templates best resonated with our audience, and identify which elements were essential to include in these materials. That led to the challenge: how could we apply this content format to the more than 2,000 topics available on the AprendiZAP Platform? If you thought, “that’s where AI comes in,” you nailed the guess!
2. Teachers and Artificial Intelligence
It was February 2023 when Débora, AprendiZAP’s product manager, invited me to a meeting about the results of the active methodologies project. Interviews with teachers indicated that students showed greater interest when they used the materials in class. We discussed what we had learned and shared this challenge: how could we scale the model we had validated with teachers to all AprendiZAP content? That was when she came up with the idea: “what if we use artificial intelligence?” Many questions about how we would do this arose, but the crucial point we defined that day was fundamental to the entire project: we would use AI with a critical review by specialist teachers.
Prompt study
We began investigating how the project could be developed. In this process, we estimated that without using AI, it would take us one to two years to produce at least one material using an active methodology for each of the 1,428 topics available in our Middle School content. Our studies indicated that with AI, we could create eight materials for each topic within three months (in other words, we would have 8,568 pieces of content available for teachers to use in the classroom).
Our development team was already investigating different artificial intelligence models, and we began conducting our first prompt tests, determining which elements were necessary for AI to create materials as close as possible to the model we had defined.
After several tests and a few weeks of investigation, we defined the prompt for generating the first thousand materials using AI, based fundamentally on the following model:
Topic: the topic of the materials
BNCC skill: the skill that will be developed
Active Methodology: the methodology selected for the topic
Curricular component: the curricular component to be addressed
Template: graphic material for the teacher to print and facilitate implementation of the methodology
Introduction: a brief explanation of the subject and presentation of the methodology
Methodology implementation stages: step-by-step instructions for implementing the methodology in the classroom
Pedagogical objectives: the pedagogical intentions teachers need to keep in mind while implementing the methodology
Assessment criteria: what will be assessed throughout the activity
Teacher’s role: how the teacher should prepare to conduct the activity
Student’s role: what is expected of the student
During the first prompt tests, we already had in mind that the materials would only be complete if we incorporated feedback from specialist teachers throughout the process.
Teachers working with AI
In April 2023, we brought together a group of 30 teachers with experience in public schools and active methodologies to join our content team. We explained the proposal and objectives, as well as our motivation for using AI in this project, and shared the fundamental role they would play: they would be responsible for the critical analysis of the materials and for providing feedback on the prompt, while making it clear that this would be a collective creation project in which we would need their input to adjust the prompt and create increasingly better materials.
Throughout the weeks of content review, we held training sessions and educational development meetings for the teaching team on Active Methodologies, Artificial Intelligence in Education, Plain Language, and Best Practices for Content Review. These collective training spaces allowed teachers to share their questions, suggestions, perceptions, and contributions so that we could create materials that truly made sense for our audience. In this process, we benefited from the experience of two teacher coordinators, Bianca Ferreira and Beatriz Bispo, who were responsible for leading and supporting the teaching team throughout the weeks of project development.
With each material reviewed, teachers indicated the level of intervention required in the content. Through this feedback, we were able to identify interesting points about the use of AI, ranging from technical to pedagogical aspects:
Technical
the limitations of AI in some methodologies and curricular components: in some subjects, AI did not have sufficient knowledge to create the materials;
the “temperature” of the prompt for each component: in some cases, AI produced very rigid materials, and we needed to adjust the “temperature” to allow for greater creative freedom.
Pedagogical
active methodology and the curricular component: in some cases, it did not make sense to implement the active methodology for the topic. For example, using the Critical Debate Model methodology for “the bisector of an angle” in Mathematics;
tips for each topic: we chose to include a tips section as a space for teacher-to-teacher communication, where our team provided recommendations for supplementary materials. Therefore, this section was not created by AI.
Based on this collaborative work model, we created an environment in which teachers shared their perceptions, played a substantial role in the use of AI, and deepened their knowledge of Active Methodologies applicable in public schools. At the end of the three-month process, they shared that they felt more confident about using active methodologies and more prepared to use AI in their teaching practices, reducing the gaps between fear and the unknown and taking advantage of the benefits this technology can offer Brazilian public education.
3. Technology in Education
The success of this project resulted from the alignment of the entire AprendiZAP team with the project’s proposal and from the contribution of each team to its development (on Fundação 1Bi’s social media channels, you can find each department’s account and learn how this experience unfolded from different perspectives).
We had opportunities for interaction between the product, content, design, development, and data teams and the specialist teachers, creating a co-creation space in which technology and education merged, with each sphere inseparable from the other. As our work became interconnected, we smoothed out rough edges, made adjustments, and redesigned elements according to the needs of each department.
MACMA 3.5 became a portrait of our work at AprendiZAP and represents one of Fundação 1Bi’s main pillars: bringing technology and education together to promote positive social impact and contribute to building quality public education.
