Independent schools across Australia are collaborating and forming networks to share the ways generative AI can be used for teaching and learning, say leaders in the sector. Widely considered the most significant educational technology development since the launch of the internet, the 2022 introduction of OpenAI’s generative artificial intelligence chatbot ChatGPT sent shockwaves around the world and left schools everywhere coming to grips with the technology’s vast potential.
In Australia, the federal government last year issued a comprehensive framework of guidelines for the “responsible and ethical use of generative AI tools in schools”; the South Australian government has launched its own custom-built teaching and learning “EdChat” AI chatbot powered by Microsoft’s Copilot, and peak industry body Independent Schools Australia (ISA) has pushed for more generative AI school assistance.
ISA CEO Graham Catt says individual Australian schools have had very different responses to generative AI in past months, ranging from a “very cautious wait and see” to an enthusiastic embrace of the technology’s potential.
“The key messages we’ve heard from teachers have been ‘we know there’s risk, we know there’s opportunity, we feel as though we’re operating on the frontier without clear guidance how we manage those things’,” he says. However, schools have also indicated the federal government’s comprehensive framework has been useful to help them balance the risk and opportunity of the technology.
The immense power of generative AI can be used by students to generate text, images and video of all kinds and many schools are concerned about the risks associated with the technology’s potential for cheating as well as identity and copyright theft.
Asked how useful generative AI was in Australian schools, AI chatbot Copilot took about a second to come up with this sensible answer: “AI has the potential to revolutionise education in Australia, but it requires thoughtful implementation, ongoing evaluation, and a commitment to ethical practices.”
Generative AI has been more challenging for less wealthy schools with fewer digital resources and there is concern it could exacerbate the existing digital divide, but Catt says there has been a “quite organic response” from schools, and many are collaborating on ways to make the most of the technology.
“Schools are forming networks and working with others,” he says. “They have formed cross-sectoral networks and are exploring and sharing examples of how AI can be harnessed. The technology is developing so quickly; there’s need for collaboration.”
Rather than waiting for governments or third parties to establish gen-AI resource banks, schools are exchanging practices and sharing resources, Catt adds, and some staff and students are collaborating on building generative AI skills, such as writing prompts, and education on the importance of verifying data and determining its origins.
“The teaching of critical thinking and ethics in that context is an important part of how educators are approaching this,” he says.
Dr Walter Barbieri, a senior lecturer at the University of Adelaide with a particular interest in educational technology, says generative AI is a massive leap forward in technology. Already a double-edged sword combining new risks and opportunities, generative AI will certainly develop further. “We are just at the beginning,” he says. “It has a long way to play out.”
Australian schools in general, though, have been better prepared for the arrival of generative AI than schools in most other countries, Barbieri adds, noting that Australian classrooms are among the most digitally connected in the world with an extremely high device-per-capita ratio. And yet there’s clearly a divide between teachers who have been in the profession for a long time and who may be struggling to come to grips with a new and very rapidly emerging technology, he says, and younger teachers who are more digitally native and more adept at working with new technologies.
Generative AI has the potential to cut teachers’ onerous administrative and lesson preparation workload, Barbieri adds, which is a “chronic and serious problem in the profession”. The AI technology can be used to assist in producing a lesson or unit plan which is part of the teaching workload, he says. A sample text could be generated using AI then tweaked by the teacher. If a slab of mock data is needed for a lesson in which students are asked to find patterns or solve a problem, generative AI can provide it.
Assessment design is increasingly critical for determining the quality of a student’s work, Barbieri says, and cheat-detecting software has proved to be “unacceptably imperfect”.