Effective Learning with Personalisation: Generative AI Adoption
Introduction
When ChatGPT was born, I saw many news and posts talking about how disruptive this would be for the education industry. ChatGPT could easily fool a law school examiner thinking the paper was done by an actual law student. What if students used ChatGPT to write essay, to cheat? Then they wouldn't actually absorb the knowledge that their teachers wanted to impart to them.
Although yes, students might resort to ChatGPT or other generative AI to get an easy solution to their coursework and homework and exams, I don't think it is necessary for us to be scared of the negative impact that AI will bring to us. In fact, any breaking changes in human history, or we call them the singularity points, always stimulated a massive changes of infrastructure of some traditional systems, revolutionised how the systems interact, and challenged the norms and status quo. Therefore, instead of trying to avoid it which is not practical as generative AI will only be more not less, we should try to embrace it, utilise it, and adapt our learning, teaching system to have a symbiotic relationship with generative AI instead.
Traditional English Learning
How do we learn English traditionally? If we try to recall, we could start from vocabulary, together with knowledge of grammar, we could start forming simple English sentences with the vocabulary that we learnt. Then, we touch on more advanced vocabulary and grammar, through textbooks which provide us some situational writings for us to digest and follow. We also learnt English through different senses, we could have audio lessons, presenting to us a scenario and the conversations happened in that scenario. We could have video lessons, providing us interesting visuals that enhance the learning experience on top of audio.
Traditional ways of learning a language, for example, English, are established for a long time and proven to be effective. But it is never personalised. We buy the same textbooks, we practice the same situational writings, we watch the same videos happening at the airport when two friends bump into each other and greet each other, or in a restaurant where our protagonist tries to order from the menu and asks where the toilet is. Scenarios are good way to get the actual language sense that helps the learner with language learning. But it can be better.
AI-Tailored Personalised English Learning
With the power of generative AI, it is entirely possible for us to design the scenarios such that they are unique and personalised pertaining to the learner's individual background.
For instance, we can have this platform where the learner will first take an introductory self assessment for the system to assess the learner's proficiency level and some of the preferences. The smart AI system can recognise if the learner is stronger in visual or audio, is a fast learner or slow learner, etc. Then the learner can further fine-tune the personalised settings to let the system understand more, for example, what is the preferred learning goal, verbal communications, or vocabulary, or professional English practices, and so on.
Another great example that generative AI can help in English learning, is having a personalised conversation with the learner. Instead of the learner watching a video where two people having a conversation with each other in a pre-defined setting, for example, at the airport, at work, at a restaurant, we can have the learner tell the AI which kind of situation the learner is interested in to practice. And the, our generative AI will simulate the situation and come out with a short scripts where the learner, very importantly, can personally participate into the conversation, with the conversational AI like ChatGPT. Furthermore, during the conversation, the AI system can recognise the learner's input, correct the mistakes made along the way, and give the learner feedbacks at the end of the conversation. For example, AI can feedback what other ways can be used to express the same idea, which tenses were used incorrectly and how should they be used in English. Or the AI could even point out whether the tone that the learner used fits the current scenario or not. In a professional setting, maybe the tone was too casual and how the learner could adjust the way of speaking to make it sound more professional.
You see, with the help of generative AI, we see more doors opened in front of us and there is endless possibility for us to design effective language learning materials. And don't forget that throughout the process, the learners will feel personally involved, the materials given to them are no longer static and permanent and all the same, they are adjusted based on the feedback loop in the learning system towards the learner's preferences and learning habits. The learners feel attached to the learning courses instead of having to force themselves to learn from some materials that they are not interested in. Therefore, the effectiveness of language learning could be greatly improved!
Challenges in Implementation
But of course, all these hypotheses are not easy to implement. Even with access to generative AI models, having a system smart enough to capture each learner's learning behaviours and build a personalised profile, could face many difficulties.
User's privacy is one concern. Some learners may not want to share too much personal information. In that case, how can our system circumvent the lack of learner's data but still try to provide a more personalised learning experience? (perhaps through the learner's interaction during a course?) If each learner has a personal profile that is constantly increasing in size as the learner progresses further in the courses, how can we handle the massive amount of data in our system and make sure the integrity of the data is not compromised? How do we make sure that our AI always generate correct contents to the learner and does not make mistakes to provide irrelevant, or worse, wrong information to the learners so that they don't learn wrong things?
All these challenges must be addressed in order for our personalised learning system to succeed in the future. And this requires a lot of research, a lot of training and practices, trials and errors, in order to fine tune the system to the ideal state that we wish to see.
Conclusion and Future Prospects
But anyway, this sounds like a very exciting journey, and I hope together with AI and its further advancement in the future for sure, we can see a revolutionised learning system that is highly personalised to cater to the needs of all different learner profiles!
0
0