AI will be used to assess student reading – 12/29/2023 – Education

Artificial intelligence, which is gaining more and more space in education and is even used to correct essays, will be tested to assess students’ reading fluency.

In 2024, an application that records students reading aloud a series of words and a short text will begin to be tested in Brazilian education networks, to assess, through technology, reading fluency. The tool was developed by a working group of Aliança Pela Alfabetização, a program of NGOs in the field of education (Fundação Lemann, Instituto Natura, Co-Impact and Associação Bem Comum) that works in collaboration with public school networks of 17 States.

Currently, an application is already being used on these networks to record the reading of students in the 2nd year of elementary school. The correction, however, is human, carried out by literacy program evaluators.

In the next academic year, a pilot correction project using artificial intelligence will begin as a way to reduce costs and speed up feedback to education networks. In developing the tool, the working group, in addition to considering pedagogical issues of reading fluency, also fed the AI ​​to take accents into account, with audio recordings from different regions of the country.

The program’s mission is to provide literacy, by 2030, to all children in Brazilian public schools at the correct stage, the 2nd year of elementary school – post-pandemic and post-Bolsonaro, the non-literate rate in this grade reached 70%.

Another tool with artificial intelligence that should begin to be introduced in Brazilian public schools next year is Khanmigo, a robot that develops a conversation with students in order to help them reach an answer.

Khanmigo was developed by Khan Academy, a non-profit organization founded in the USA, which offers a free teaching platform, and by OpenIA, the startup responsible for creating the famous artificial intelligence ChatGPT.

The tool was brought to Brazil at the beginning of 2023 by the Lemann Foundation, to be tested with a group of 200 students, in a pilot project. The institution intends to expand this test to 10 thousand students by 2025 and, subsequently, implement it on a large scale in public education networks, on a partnership basis.

In addition to helping students with their doubts, Khanmigo can also be used by teachers to prepare lesson plans.

The Lemann Foundation also provided financial support to two other AI initiatives for education, Letrus, a writing correction platform, and the Associação Nova Escola project to develop free lesson plans for teachers.

In May 2023, Lemann brought to Brazil Sam Altman, the founder of OpenIA, who spoke enthusiastically about the potential of AI for education, which could, in his vision, give “each student a personalized teacher”.

Since then, the impact of AI in schools has only increased, and views on this process range from optimistic to catastrophic – we had throughout the year, for example, the “boom” of cases of boys using AI to make fake nudes with girls face.

Executive director of the Lemann Foundation, Denis Mizne, 47, told Sheet that his view on the use of AI in education is “moderately optimistic.”

“The risks are not negligible. It’s a black box, not even the owners themselves know exactly how it works, and this creates anguish,” he said.

However, he is not afraid of using it, much touted, for school work. “This seems similar to the time when Google emerged, and the fear came that students weren’t going to study anymore, produce anything, they were going to ask Google for everything,” he said.

“When I was a law student, many years ago, there was a collection of summaries of legal books. Some students avoided reading the books in full, but then the consequences came”, he compared. “As a society, we have more and more access to content, and obviously ChatGPT is a huge leap forward in relation to the collection of summaries and even Google. But this forces education to be a place of critical thinking, of building synapses, of analysis.”

For him, with this advancement in artificial intelligence, “the importance of students expressing themselves with their own ideas” must grow in education. “Learning content has to be at the service of analysis, not memorization”, he argued.

Mizne says he believes that, with AI, “we can free class time from content that is boring, because it is already available to students.”

He is also optimistic about the possibilities that AI offers for personalizing learning. “Students from expensive schools hire private teachers when necessary. Those from public schools cannot do this, and inequality increases”, he pondered. “AI must create the possibility of unlimited and free quality tutoring for public school students.”

Of course, there is another side to this story, Mizne highlighted, citing fake nudes in schools, the destruction of reputation, among other damages. “This is all very serious,” she said. “We are seeing several revelations of big tech’s complacency towards children and adolescents on social networks. We already have a huge amount of information about the damage this has to mental health, especially for girls.”

Mizne advocated for the regulation of technology companies. “We entered a bubble where we were convinced that the best regulators of technology are the companies themselves. And that cannot happen, the State needs to regulate and soon,” he said.

Furthermore, he stated, we need to invest in media education, that is, “educating children, young people and families to separate the wheat from the chaff” in the digital universe.