Asia leading charge on AI ELT research
The continent of Asia is at the forefront of research into AI in English language teaching, according to a landmark report from the British Council.
Artificial intelligence and English language teaching: Preparing for the future, surveys teachers, shows viewpoints of expert witnesses and examines research to see how the sector can effectively interact with AI down the line.
In terms of literature, 19% of articles on AI in ELT have come from China, with 72% coming from Asia as a continent.
“It’s a massive swing. The superpowers in AI are the US, China and Europe, but the latter is probably a fair bit behind those two.
“That swing to China specifically, for policymakers, especially in the US and the UK, is quite a big thing, really. It highlights the amount of investment that the Chinese are putting into it,” Adam Edmett, one of the report’s authors and the British Council’s head of edtech innovation told The PIE News.
Another point from research across the decade that the report reviewed showed that, despite the possible capabilities, listening is not an area where AI is being used for support.
“One look at audio-visual methodologies employed by VR companies should be enough to show that we have a long way to go yet,” said Gavin Dudeney, director of tech at The Consultants-E, who was an expert witness in the report.
The development of speaking, writing and reading skills were much more prevalent in the use of AI – as well as supporting pedagogy and self-regulation, the report said.
“But you can see why that is – that listening is nonexistent in terms of AI adaptation.
“In classrooms around the world, it’s not unusual for a teacher to have a language level of like A1 or A2, and they have a large class, maybe 50 students. When you look at that, in reality it’s really difficult for the students to practice.
“The teacher in that situation also can’t give that individual kind of practice – AI isn’t really, at the moment, in that form to offer excellent listening practice,” Edmett explained.
Limited capabilities, incidentally, was one of the challenges of AI that was listed by the report – alongside technology breakdowns, especially through poor connectivity – something Edmett’s team has been working to overcome.
Edmett’s team won a bronze award at the Learning Techonlogies Awards in November for their efforts in offline app development for learning in Nigeria and work on a learning foundations program in Rwanda.
Other challenges AI in ELT has come across is a familiar one – fear – which included lack of clarity on the storage of information, uncertainty of the technology’s operation, and a usual suspect in this space: fear of the unknown.
In terms of how teachers use AI in their teaching, creating materials was by far the most popular use, with 57% listing it. Some 53% use it to help learners practise English, and 43% use it to create lesson plans.
An overwhelming number of teachers also said they either agreed or strongly agreed that AI could help learners improve their English skills in all four methods -76% said it would help with speaking, 75% said it would help with writing.
Despite the little use of it in classrooms according to the research for listening, 74% still agreed it could help learners improve in that area, and 79% said it would help with reading.
On whether AI could have a negative impact on students’ ability to improve their English, it was quite an even split – 36% agreed or strongly agreed, while 34% disagreed or strongly disagreed – another 30% were neutral, suggesting there are mixed feelings all around.
Most agreed that learners should be able to write in English without the help of AI tools.
The problem, however, according to Edmett, is that AI could mean just about anything – just as digital meant just about anything 10 years ago.
“It seemed to be the solution to everything – people would say, ‘what we need is a digital solution’. And that mentality is now being taken over with AI.
“There were things that were just described as digital tools like Grammarly and spellcheck – and it’s now being described as AI because people want to get on the bandwagon,” Edmett explained.
In terms of the regulatory framework for AI in English language teaching, experts said that while they were emerging, no global consensus was formed yet – and that perhaps a reset was needed. Most of them said tech companies shouldn’t be left to regulate themselves.
“I don’t think [owners of the platforms] are inherently evil – but the tools which they generate can certainly be put to that use.
“I’m sure they don’t have education at their heart either ,despite what they say,” said Thom Kiddle, director of NILE, in the report.
The report suggests that, as long as definitions can be codified and a set of principles can be agreed upon, ethical concerns can indeed be tackled and help AI improve, “and not hinder the learning process”.
“Whether new technologies will bring widespread systemic change that matches the AI hype is an ongoing debate,” the report read.
“A reading of the history of education technology would say otherwise.”