The UK government turned to AI to predict students’ grades during the pandemic, when students couldn’t complete exams in person. But the algorithm was scrapped after it disproportionately marked down students from disadvantaged backgrounds, wrote Forbes.
Creating strong regulations and ensuring teachers are on-board are two steps that schools can take to ensure classrooms adopt AI smoothly. Addressing these concerns will help schools realise AI’s potential for personalising learning.
At the recent AI x GOV summit, public sector officials and education experts gathered to discuss the responsible and ethical adoption of AI in schools, and what this technology can bring to the learning environment.
Recognising the benefits of AI
The panel ‘A classroom of one: Personalising the student experience with AI’ explored three ways the tech can benefit the classroom.
First, AI can help enable personalised learning, highlighted A/Professor Wenli Chen from Singapore’s National Institute of Education. Algorithms can analyse how students fare across different online learning materials, identifying individual areas for improvement, she explained.
This helps teachers to understand each student’s individual needs. Currently, teachers face larger class sizes and limited time to attend to unique needs, she cited from her research into learning sciences.
She also highlighted Singapore’s National AI Strategy, which outlined plans for a AI-enabled marking system which assesses students’ essays and written responses. It is also looking to build an AI chatbot that will motivate students and recommend further learning tasks.
Second, AI of the future may be student-led, rather than a system adopted by the school. Students could choose to use AI on individual projects they’re working on, for example using AI to help writing poetry, shared Anna Korpi, Counsellor Education and Science, Embassy of Finland, Singapore.
Third, AI can help analyse data to understand student “metacognition”, which looks at how students learn how to learn, shared Ibrahim Bashir, Technical Project Manager, Educate Ventures Research.
An EdTech company was able to quantify this metacognition, highlighting that positive feedback and practice were important to doing better in assessments. Seeing this laid out in statistics is the first step to producing new ways for students to learn more effectively, Bashir continued.
Transparent and explainable AI
To realise these benefits of AI adoption, schools will first need AI tools to be explainable and transparent. Sometimes these tools are “built by some computer scientist, but without input from school personnel or teachers at all”, shared Chen.
This means as educators “we have no idea where their data is coming from, where they are building such algorithms”, she explained. Without oversight or transparency, AI may analyse data out of context or incorrectly – something that worries teachers, she added.
The first step to creating explainable AI is to regulate. Korpi highlighted that Europe has strong personal privacy laws. Individuals need to be aware of what data is being collected and how it is used.
Even one negative incident in this area “will really hinder the adoption” of AI and technology in the classroom, Chen said. Governments can create frameworks around AI and data, focusing on maintaining ethical decision-making and protecting data privacy.
Another way to ensure responsible use of AI is to recognise the local context in schools. For example, schools may not have historical data about a student, affecting the accuracy of AI tools. In this situation, it’s important to proceed on a case by case basis, Bashir said.
“There is no substitute for local context”, he summarised. Bashir highlighted that schools should recognise that AI is a sensitive topic, with parents reacting to data use differently in different parts of the world.
Addressing teachers’ concerns
Being replaced is the most immediate concern that some teachers may have regarding AI. But no matter how smart the technology is, “it will never replace the teacher”, Korpi emphasised. The technology should be seen as an enabler or assistant, she added.
This is especially the case during the pandemic, where human interactions have become a luxury, Korpi shared. These human interactions are important as learning remains a social process between people, she explained.
Chen agreed that teachers won’t be replaced by AI. The technology excels at processing large amounts of data, identifying patterns, and taking over repetitive tasks. But there are some important areas where humans excel, such as understanding situational context, she shared.
Teachers may also feel some hesitation towards adopting new tools. Their concern stems from the question “What value will new technologies bring to my students?”, shared Dr Wagheeh Shukry Hassan, Deputy Director – Educational Technology & Resources Division, Ministry of Education, Malaysia.
The ministry now looks to answer that question. One way that AI can create value is by highlighting to teachers how that subject may be taught in different ways, or what types of content students will enjoy learning about, he shared.
Ensuring that AI is transparent and demonstrating to teachers the potential of the technology are hurdles facing schools. But while challenges lay ahead, they shouldn’t “paralyse us” and stop the development of new educational technologies, said Korpi.