In this exclusive interview, Fabio Cerpelloni sits down with Sophia Mavridi, a digital learning consultant, teacher educator, and senior lecturer specialising in educational technology, to discuss the role of AI in language education.
Drawing on her research and classroom experience, Sophia highlights the importance of critical engagement with AI tools, the need to educate teachers – not just students – about technology, and the challenges and opportunities that come with integrating AI meaningfully into teaching and learning.
Fabio Cerpelloni (FC): How can AI help students become more independent in their learning?
Sophia Mavridi (SM): There are lots of ways this can happen. Let’s take writing as an example. AI tools can give feedback on grammar, vocabulary, coherence, and tone, which are all areas where language learners typically need support. But just having access to these technologies doesn’t mean students know how to use them effectively to improve their writing skills.
Students need a certain level of feedback literacy, which is the ability to interpret, evaluate, and apply feedback. Without this, interacting with an AI tool might confuse students, not least because these tools have their flaws. They make mistakes. And by “mistakes” I don’t mean glitches or bugs. I’m talking about a fundamental feature of how this technology works.
AI is not intelligent in the human sense. It doesn’t understand meaning or context. It doesn’t engage with emotions or cultural significance. It generates language based on statistical patterns. This is something teachers and students should know.
FC: What strategies can educators employ to guide students in using AI responsibly?
SM: We often place the responsibility on the teacher. We say it’s the teacher’s job to integrate innovation and educate students on how to use technology responsibly. But who’s educating the teachers? Are they being prepared for this?
Let me bring up the pandemic as an example. When language teachers were suddenly thrown into remote teaching—many with little digital background—they were expected to build the plane while flying it. And they did—with resilience and a lot of informal learning. But the pandemic also exposed how unprepared the system was.
Now the trend is AI. Rather than, by default, expecting teachers to educate students on how to use this technology, I’d flip it: educational systems must first educate the teachers. And I don’t mean just training them to use Tool A or Tool B. Tools are useful, yes—but that’s not enough. We need digital pedagogy. That means understanding not just how to use tools, but why – linking them to learning objectives, theories of learning, and students’ needs.
There’s a big difference between using a tool and teaching well with technology.
FC: Do you think new training courses should be offered on digital pedagogy?
SM: Absolutely. Educational systems – including conferences, MA programmes, undergraduate degrees – need to prioritise this. Too often, conferences seem to focus on tools, not pedagogy.
I published research[1] into postgraduate programmes and explored whether and how they develop language teachers’ digital expertise. The findings show there’s still a lot of room for improvement. Degree and language certification programs need a component not just on how to use tech, but on when and why to use it.
We need to ask: What’s the purpose of using technology? Are we enhancing students’ learning through it?
Take Kahoot, for example. It’s a good tool. But if we only use it for multiple choice, that’s just a digital version of a paper quiz. Fun and learning are related—fun can support motivation—but fun on its own doesn’t necessarily lead to learning.
FC: How do you use AI in your classes?
SM: I educate students about what AI is and what it isn’t, how it can support learning, and how it might bypass learning if used uncritically.
These conversations aren’t one-off lessons. Like all digital literacies, this needs to be embedded in our lessons, discussions, and reflections. With our students, we need to talk about opportunities but also the issues that come with AI, like transparency, labour exploitation, and environmental concerns.
And we need to be honest. Institutions say, “embrace innovation,[2] ” but then penalise students for using AI. That’s confusing. Instead, we should educate students and then invite them to make informed decisions and take responsibility.
As for my own use, I use AI to develop teaching materials and lesson ideas. I tell my students how I use it, so they can explore ways they might want to use it too.
FC: What did your own research into AI reveal?
SM: I’d like to point out that my research wasn’t hype driven. It began in 2022, a time when AI wasn’t such a hot topic like it is today.
I explored students’ experiences with AI, from the students’ perspective. I didn’t just explore what they did with AI, but also how they navigated the experience, how they negotiated the opportunities and limitations, and how they defined those. It was fascinating to speak with them.
What really struck me was how sophisticated many of their uses already were. I found that many students used AI to support their learning, not bypass it. All these uses are ad hoc, unsupervised, and sometimes even done in secret, especially in institutions where AI is banned. And I’m not talking about copying and pasting AI-generated content. The students themselves said that simply copying and pasting isn’t something they value or accept as good practice.
Instead, I’m talking about things like paraphrasing, editing, and refining—actions that might lead some educators to look at students’ work and ask: Is this the student’s work or the machine’s? And what learning might have been lost in the process?
From the teacher’s perspective, it might look like students are trying to hack their way to success. But when the system is about marks and only about marks, some students might want to use AI to get higher scores.
Students are aware of academic integrity issues, too. It is a myth that only teachers and institutions care about these. Students do care.
So, my research led me to ask some important questions: Is AI the problem? Or is AI simply breaking what is already broken?
Take writing as an example. Many of the tasks we assign are still formulaic—twentieth century-style writing tasks, like the five-paragraph essay. Generative AI is extremely good at producing those.
So, what’s the real issue? AI or the task itself?
Biography
Sophia Mavridi is a senior lecturer in educational technology and TESOL at De Montfort University, UK, where she teaches on the MA in ELT and leads digital learning in her department. She is also a teacher educator who has worked with major organisations such as the British Council, NILE, Bell Foundation, and International House. Her publications reflect a strong interest in innovative pedagogies in language education.





