AI can help us find some answers in the elusive quest for ‘quality’ in education

19 Mar 2023

Even the talking bot ChatGPT knows it is wrong to confuse compliance with evaluation when trying to assess how schools fare in providing quality education for all children in South Africa.

I came across an Amnesty International report published in 2020 titled Broken and Unequal: The State of Education in South Africa. It argued: “South Africa is failing too many of its young people when it comes to education. Although it has made significant progress since the end of apartheid in widening access, this has not always translated into a quality education for all pupils.”

The report noted that the system continues to be dogged by stark inequalities and chronic underperformance that have deep roots in the legacy of apartheid. However, Amnesty International argues that this is still not being effectively tackled by the current government, and the result is that there are still many schools with crumbling infrastructure, overcrowded classrooms and poor educational outcomes.

I decided to embark on an experiment using artificial intelligence (AI) in the form of ChatGPT to search for answers. We’ve been hearing a lot about AI lately and its ability to generate graduate-level essays, solve complex mathematical problems or even set examination questions with unbelievable alacrity.

In reviewing the debate around quality education globally, ChatGPT immediately went to the root of the question of what constitutes high-quality education and how to ensure that all students have access to it. It noted the growing recognition that quality education is critical for the development of individuals, communities and nations, but, equally, the great disparity in the quality of education provided throughout the world.

“One of the main issues in the quality education debate is the unequal distribution of resources, including qualified teachers, textbooks and technology, among schools in different regions and countries. Many argue that addressing these resource gaps is a key step towards achieving quality education for all,” ChatGPT said.

It then moved on to other issues such as the relevance of education and the need for education systems to prepare students for the challenges and opportunities of the 21st century, including the increasing demands of the globalised economy, the need for digital literacy and the importance of sustainable development.

ChatGPT ended the discussion with a little summary, which seems to be typical in these types of inquiries (and a possible clue for teachers who want to know if students are using AI):

“The quality education debate is complex and multifaceted, with many stakeholders and perspectives involved. However, it is widely agreed that improving the quality of education is essential for promoting social and economic development, reducing inequality, and creating a more just and sustainable world.”

The Amnesty International report reminds us that the state of education in South Africa must be seen within the wider context of one of the most socioeconomically unequal countries in the world. Black South African households earn on average less than 20% of white households, and nearly half of the black population is considered to live below the poverty line, compared with less than 1% of the white community.

In other words, the child’s experience of education in South Africa still very much depends on where they are born, how wealthy they are and the colour of their skin.

Visit Daily Maverick's home page for more news, analysis and investigations

Compliance versus evaluation

We are trapped in a compliance versus evaluation quagmire. ChatGPT points out that “compliance and evaluation are two different approaches to school quality that serve distinct purposes. Compliance refers to ensuring that schools meet certain legal or regulatory requirements, while evaluation is focused on assessing the quality and effectiveness of a school’s programmes and practices.”

The truth of the matter in our country is that we have never really been able to move beyond compliance and implement measures to improve school quality.

Umalusi evaluations, whole-school evaluations and integrated quality management systems (IQMS) consume a lot of time at school level and, even if their intention is to improve quality, they have generally not been able to escape from the compliance trap. Compliance assessments often involve checking to see if schools are meeting specific criteria or standards, and the focus is on ensuring that schools are meeting minimum requirements rather than improving overall performance.

Even ChatGPT knows that “evaluation, on the other hand, is focused on assessing the quality and effectiveness of a school’s programmes and practices. Evaluations may involve gathering data on student outcomes, teacher effectiveness, curriculum implementation and other factors that influence student achievement. The goal of evaluation is to provide feedback to schools and educators that can be used to improve performance and outcomes for students.”

ChatGPT provides a way out of the quagmire, saying “compliance should not be the sole focus of school assessment, as improving overall performance and outcomes for students requires a more comprehensive approach that includes evaluation and feedback”.

In concluding, ChatGPT keeps its answer on the general level and cleverly says that “education quality is a complex issue that is affected by multiple factors, including funding, infrastructure, teacher training and support, curriculum development, and societal factors such as poverty and inequality”.

ChatGPT provides some valuable advice to the government on how to evaluate schools. It reviews and observes some of the following trends globally:

Focus on equity: There is a growing focus on equity in school evaluation. This includes evaluating how well schools are serving learners from historically marginalised groups and taking steps to address disparities in student outcomes.

Emphasis on social-emotional learning: Schools are placing a greater emphasis on social-emotional learning and evaluating how well they are promoting positive learner outcomes in this area. This includes assessing learners’ ability to manage emotions, build relationships and make responsible decisions.

Use of data: There is an increasing trend towards using data analytics to evaluate school performance. This includes using data to track learners’ progress, identify areas for improvement and measure the effectiveness of interventions.

Standards-based grading: Schools are moving towards standards-based grading or assessment, which assesses learners’ mastery of specific learning objectives rather than just their overall performance. This approach provides detailed information about what learners know and can do, which helps teachers to better tailor instruction or teaching and interventions to their learners’ needs.

Though I don’t know how it really works, I assume that ChatGPT is reflecting back the general themes of school quality and identifying the gaps in implementing strategies.

An experiment like this raises many questions. What does it mean for a computer to assess quality, or to describe the elusive broad evaluative ideas? Is AI suggesting anything better than what I might have heard or said? Can a meaningful sense of quality really be artificially generated? More importantly, can AI improve the quality of schooling for real children in classrooms?

The answers provided in this experiment do resonate with my own professional experience, and perhaps, with more time, better approaches can be found to improve schools using AI. DM168

Dr Mark Potterton is the principal of the primary school at Sacred Heart College and director of the Three2Six Refugee Children’s Education Project.

This story first appeared in our weekly Daily Maverick 168 newspaper, which is available countrywide for R25.

Categories:

Tags:

DM is much better if you are signed in...

South Africa

Compliance versus evaluation