The Impact of Artificial Intelligence on Learning in Brazil - AI-powered writing feedback study

Sep 6, 2026

The report evaluates how artificial intelligence changes learning outcomes in Brazilian secondary schools by using automated writing evaluation (AWE) systems to score and comment on essays. Researchers examine whether AI improves students’ writing skills and exam-related performance while reducing teacher time constraints in grading. In schools implementing AI, teachers provided more frequent individualized feedback, and students’ essay scores increased. The evaluation is grounded in the Post-Primary Education Initiative (PPE) and an AEA RCT registration record (AEARCTR-0003729).

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages4

Reading Time10 min

AccessFree to download

Policy issue

Improving learning in low- and middle-income countries

Context of the evaluation

Literacy and writing goals in Brazil

ENEM and the writing portion

Technology access in low-connectivity settings

Details of the intervention

Study setting and sample

Enhanced AWE intervention

Pure AWE intervention

The Impact of Artificial Intelligence on Learning in Brazil Researchers: Bruno Ferman Lycia Lima Flávio Riva Sector(s): Education, Artificial Intelligence (AI) Location: Espírito Santo, Brazil Sample: 178 public schools encompassing ~19,000 senior high school students Initiative(s): Post-Primary Education Initiative (PPE) Target group: Secondary schools Students Teachers Outcome of interest: Student learning Technology adoption Take-up of program/social service/healthy behavior Intervention type: Digital and mobile Online learning Pedagogical innovation Technology AEA RCT registration number: AEARCTR-0003729 Partner organization(s): Espírito Santo's Education Department (SEDU/ES), Lemann Foundation, [Google.org](Google.org), Oppen Social, Fundação CAEd Artificial intelligence (AI) has the potential to support teachers in completing time-intensive, subjective tasks, such as grading essays, and to provide individualized feedback to students on their writing. However, evidence on whether and how AI impacts students’ writing skills remains limited. To fill this gap, researchers evaluated the learning impacts of using AI systems to score and comment on essays written for Brazil’s national post-secondary admission exam. In schools where AI technology was introduced, teachers were able to provide more frequent individualized feedback to students, and students’ essay scores improved as a result. Policy issue Improving learning is one of the most pressing goals for educational policy in low-and middle-income countries. A common barrier to learning is that teachers are time-constrained and often have to juggle between providing individualized assistance to students and performing routine tasks, like grading. This is especially true for the development of writing skills, which requires teachers to spend long hours grading. Educational technologies (ed techs) could help alleviate this challenge by performing the operational parts of teaching and allowing teachers to reallocate time from less to more complex tasks (e.g., one-on-one feedback to students) . For example, automated writing evaluation (AWE) systems can potentially help improve writing by using natural language processing and machine learning algorithms to predict scores and allocate feedback, thereby reducing the amount of time teachers need to spend on grading. However, there is little evidence on the effectiveness of ed techs or artificial intelligence (AI) focused on writing. Are AWE systems effective in improving students’ writing skills? Context of the evaluation Providing high-quality education is a challenge in Brazil, especially when it comes to language skills. According to the 2018 PISA exam, a worldwide study of students’ scholastic performance, the average 15-year-old Brazilian student scored 413 points on reading, compared to an average of 487 points in all OECD countries.1 In response to the need for higher-quality education in Brazil, the implementing partner in this evaluation was launched in 2015 with the mission of improving literacy and writing skills among school-aged youth by applying artificial intelligence to linguistics in public schools. Its main product was a pedagogical program that provided feedback on writing skills to students, using an automated writing evaluation (AWE) system combined with validation of feedback by human essay graders. The AWE system was embedded in an online platform that granted students access to in-classroom practice opportunities for the essay of the National Secondary Education Exam (Exame Nacional do Ensino Médio, ENEM) . ENEM is the second largest college admission exam in the world. It has been increasingly used as an entry exam by many postsecondary institutions in Brazil, and the essay portion of it accounts for the largest share of the public-private school achievement gap in ENEM scores. In allowing teachers to reallocate their time from grading to providing more tailored assista

FAQs

What gap in evidence does the study address about AI in education?

The study targets limited evidence on whether and how AI affects students’ writing skills, especially through scoring and commenting on essays.

How does the AWE system change teacher and student interactions?

By generating feedback and scores digitally, it allows teachers to provide more frequent individualized comments while managing time-intensive grading tasks.

What does the evaluation measure in Brazilian public schools?

It measures impacts on students’ writing skills and essay scores, comparing schools assigned to enhanced AWE, pure AWE, and a control structure across 178 public schools.

Sep 6, 2026

Upload by Valentina