Talk by María Teresa González from IH Mexico City.
Given at ANUPI 22nd International Conference: “SHAPING THE FUTURE OF ELT: TRENDS AND INSIGHTS FROM THEORY TO PRACTICE”
I was very excited to be invited to the conference this year. It took place in one of the most beautiful beach destinations in Mexico… but no, it wasn’t Cancún! It was held in Zihuatanejo. It is a picturesque small town in the State of Guerrero, on the Pacific Coast.
ANUPI (stands for the National Association of English Teachers at University Level) is a very well-known, non-profit organisation that works to promote English teachers' professional development all over México. One of their main goals is to improve the quality of teaching English as a foreign language (EFL/ESL) in the country. Every year they organise this academic event to offer teachers a series of talks and workshops showcasing the latest language learning and acquisition trends. It was not a surprise that this year there would be many talks (I counted at least 5 in the programme!) on the use and alternatives that Artificial Intelligence has been offering in the last few years.
There were teachers from all ages and backgrounds. So, I thought I needed to start with some common ground like basic concepts of generative AI and key components of human intelligence such as learning, problem-solving, perception, reasoning and language use along with some examples of how we use AI in our everyday lives, even without noticing. And it is precisely about language use that I focused on the presentation. According to Yule (2006) humans are continuously creating new expressions and utterances manipulating linguistic resources to describe objects and situations. This is often referred to as language productivity. As opposed to birdcalls and traffic signs, full-fledged human languages can offer an unlimited variety of sentences.
Human reasoning allows us to adjust meaning to the right context to make sense. This is still one of the biggest challenges AI has yet to face. A good example of this is what happens in Large Language Models like ChatGPT because sometimes, depending on the quality of the prompt, the result can be either very effective or completely nonsensical!
There are four key principles to writing prompts:
- Be specific: as much as you can.
- Keep it simple. The more concrete the better.
- Use examples. There is a big power of association of ideas in AI, so go for it.
- Refine and iterate. It is possible that you do not get the desired response in your first interaction, but if you repeat the prompt with some adjustments, chances are you will.
However, prompts were not the central part of this talk. There are lots of webpages offering prompts already tried and refined for English teachers’ contexts. Try this website for instance.
What I wanted to show teachers was how interacting with some AIs can motivate students and at the same time help them practice or develop their language skills.
To begin with, we can have students chat with ChatGPT and ask it questions and make dialogue as if it was a real person. This is helpful because ChatGPT has a free version. It is safe to use in class. It is adaptative to the student’s level. It offers high-frequency language samples. It can correct learners’ language. Students can also receive immediate feedback on their language use. You can play 20 questions, all sorts of role-plays, mystery quests, charades, make up songs and poems, create stories and ChatGPT will be a good helper in the classroom.
After using ChatGPT in different classes and with students from different levels, it was observed that they mostly had a positive reaction. It helped them build their confidence. Others enjoyed the wide range of topics that can be discussed there. Others mentioned they felt safe and less stressed to use the language as they were using it with “someone who could correct them but was not a teacher…”. The ones who like to be corrected appreciated how in a subtle way ChatGPT can offer a better way to say something.
Another tool that students enjoy using in class is https://www.imagine.art/, here they can create images and later describe them or talk about the errors in the AI understanding of what is needed. For example, in a class about science, students got this:
Then they tried some refining and iteration and had some fun:
This is what they got when they tried more specific and enriched prompts.
Later students were asked to discuss what went wrong and what needed to be included in the prompt to generate the right image. Students were highly engaged in this activity and at the end of the session, they shared that they enjoyed the visual stimulation and the challenge of describing unusual images. They used their creativity and collaborated in group discussions and games.
If your students enjoy contests and drawing https://quickdraw.withgoogle.com/ will be a great AI to use in class. Here students have 20 seconds to draw an object. If the AI guesses what students are drawing, they win. If not, it shows them drawings that were close and what other people drew to represent that object. Students can discuss their drawings and compare them.
AI offers both teachers and students creative ways to use language. Some people might still feel AI is a big risk for education. However we can conclude that to use it effectively in class, we need to teach students how to use it efficiently and ethically. A code of ethics for using AI in class shows students that AI is an aid but that it should not be used to cheat. AI gives teachers an incredible opportunity to evolve, inspire and be inspired!
You can see Maria Teresa's presentation here.
Interested in speaking at an event but worried about the cost?
The Speak for IH grant of €300 is available to anyone currently working at an IH school who is travelling to another city to speak.
>>> Find out more about applying for the Speak for IH grant here <<<