A Tsunami of Ideas from our Readers

A Tsunami of Ideas from our Readers

By: Think Tank Readers

Editors: At our request, a number of language teachers sent us their comments on ChatGPT. Most are positive. Most include good classroom applications. All are insightful. But what else could we have expected? Anyone who reads a magazine on connecting brain studies to their profession is bound to be ahead of the herd!

Goodbye to Dry Material

Roger Blievernicht - Teacher of English for Academic Purposes

Let’s be honest, we’ve all been there—staring at a textbook, trying to power through a chapter that’s about as captivating as watching the clock tick. I mean, would I be far off in claiming that most students don’t find joy in reading about the history of the postal service in America, a biography of Michael Jackson, or conservationists in some remote part of Alaska? But here’s the thing, even the driest subjects can hide some valuable lessons—be it vocabulary, grammar, or critical thinking skills. So, what do you do? You make it interesting, now in a matter of seconds.

One of the most remarkable things I discovered about Chat GPT is that with just a few clicks, you can instantaneously whip up a reading on any topic that is… interesting! Are your students excited about how well Japan is doing in the world cup? What about creating an article on the history of Japan’s performance at previous world cups? Are some of your students going to be dressing up in kimonos and suits for the Coming-of-Age Day? Why not create a reading that explains how adulthood is celebrated around the world in a fascinating way?

One of the best parts is Chat GPT can even add humor, making the dullest subjects pleasurable to learn. It is now possible to even customize the reading by adding vocabulary words, comprehension questions, and changing the style to suit the English level of your learners.

Ladies and gentlemen, we have now entered the age where we can easily substitute tedious textbook material with an unlimited amount of content that is both academic and fun.

ChatGPT as an Assistant Analyst

Cheryl Woelk - Language and peace education coach, Collective Joy Consulting in Seoul, South Korea

Personally, I think that innovative technologies like ChatGPT are an opportunity for learning together and developing our teaching further. Regardless of the challenges, reacting in fear and worry does nothing to serve our students and only leaves them on their own to navigate the integration of these tools in their language learning.

I’ve played with ChatGPT a bit so far, and I think there is an exciting potential for its use in English classes as a language model. With high intermediate and low advanced learners, I often draw their attention to the patterns of organization in speech or written texts and prompt learners to reflect on the similarities or differences with other languages they know. I’ve found this process can help to provide a scaffolding for both receptive and productive skills, easing some of the mental load of processing ideas and organization at the same time. I emphasize that these patterns, though, are largely descriptive rather than prescriptive.

Prompting ChatGPT to identify different discourse patterns and asking it to provide examples on different topics with varying levels of detail could be an effective way to look at organizational structures together. It could also be an assignment in itself: ask ChatGPT to create a paragraph with a certain structure, then critique its level of cohesion and effectiveness and provide editing suggestions. As when working with other data sets, like corpora, we may be surprised at what we learn and observe through these types of interactive and analytical tasks.

Unexpected ChatGPT Risks

Pauline Bunce - Australian educator

I think that schools and universities will work out ways to use and manage ChatGPT. However, I worry about its use in the wider community, specifically essay- and story-writing competitions that are run by organisations that may not be fully familiar with the use of AI.

It could also be used by scammers, mimicking the language of service providers in order to persuade the unwary that they need to make payments to (false) bank accounts. At present, it is often the rather clumsy language used by such scammers that gives them away. 

Discourse through ChatGPT

Chris Clancy - Saku Chosei High School communicative English language teacher extraordinaire

ChatGPT is an amazing piece of software. I see it as a tremendous opportunity for language teaching and learning. I suggest we embrace ChatGPT as a part of the new “new” brought on by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic that has forced us to alter our lifestyles in so many ways over the past three years. As we have so often come together in times of trouble throughout history, it’s time we educators similarly overcome the dualistic mentality that plagues society as well as our profession. Be it esteemed university-type educators versus other educators, upright moral educators versus potential plagiarists or intelligent educators versus artificial intelligence, the time to coalesce is now. 

Applications such as ChatGPT are the way of the future and offer more benefits to language teaching and learning than demerits. For every gptZEROx used to gauge whether a learner’s content might be original or not, there will be newer applications to question the others’ accuracy. Educators must learn to facilitate the teaching and learning experience between learners, artificial intelligence, and themselves.

My first reaction to ChatGPT was positive. What I saw was unlimited access to discourse interaction in the target language. Second language educators all know that the same discourse interaction through which we all learn our native languages has been a largely absent element that is necessary for target language acquisition to occur. Written and aural texts have been restricted affairs in which learners are faced with preconditioned language to which they must respond. Even teachers’ bold attempts at interactive journals or face-to-face tutorials are impossible to maintain with all students. Although ChatGPT is currently only available as written discourse, it has the potential to fill a previously missing role that has prevented learners’ success in second language acquisition.

Let’s Embrace the Benefits of ChatGPT

Mirela Ramacciotti - Lawyer, teacher, author, and PhD in Neuroscience and Behavior, Brazil

This was originally a post written by invitation of LinkedIn News (see at [LINK]). It is an informed opinion on banning chatGPT from schools. And I say informed because it’s based on evidence about how people learn. It is an opinion because it comes from a personal understanding, exercised through reflection, that I committed to when I received the invitation.

Let’s start by cutting to the chase: banning this new AI tool for the production of texts. I don’t think banning is the way. First because it’s already accessible to everyone. Second because it fills a gap and exposes an undisputed truth: compiling facts, putting them in a coherent sequence, bringing consistent information in a cohesive and exponential way—without flaws in this construction—is no longer a human prerogative. Technology has reached a threshold previously thought unattainable: aligning ideas to produce a text. Taken together, both arguments suffice to bring us to what matters most: why is writing texts being taken as an essential step in learning?

And here I begin by listing some of the cognitive benefits of writing that remain unchallenged thanks to the science of learning: writing recruits a set of neural networks—these still human and in their joint recruitment still without reproducibility by any AI tool—which involve attention, memory, reflection, reasoning, decision making, adequacy of information, and critical thinking. This list of skills becomes a combined set through exercise, that is, the employment of writing in tasks that lend themselves to their recruitment. And now I ask you: have we (teachers) proposed to students themes/questions that can be answered better by the newest AI tool (for now, Chat GPT)?

If in doubt, I invite you to get to know the tool and do a brief exercise: insert your theme and compare the answer obtained with students’ answers. If the tool’s response is better, we can infer that some of the cognitive abilities listed above were not optimally recruited by the students.

ChatGPT Boon

Curtis Kelly – or is it really?

I was thinking about ChatGPT and listening to an NPR radio show about a new AI psychological counseling app, when it hit me. This new technology is going to help solve a particularly nasty problem, though not one in education. I cannot tell you how many times I have called a customer service number to ask a question about a product and fallen into this cycle of horrors:

  • I wait through an explanation of five numbers I can push, none of which quite fits my problem, and then five more, although I already know I’ll have to talk to a live agent.
  • Then I speaker phone through a two-hour wait until a live agent is free.
  • Then the agent, obviously handling six callers at once, eventually states he or she is not authorized to give me an answer, so tells me to call another number.

Site “live” help is not much better. It is usually just a bot trying to maneuver you to the same irrelevant information you saw on the FAQ page. Since I moved to the US, I have been put into customer service purgatory a few times a week. I hate it.

It will not be long though, until technologies like ChatGPT can solve this problem. You ask this much smarter bot a question, and it not only knows what you mean, but it has a whole company database of policies and previously called-in exchanges to consult and find exactly what you need.

Hallelujah! I see the light!

How Do you Solve a Problem like ChatGPT?

Matt Ehlers – One who enjoys doing web search and evaluating online information

There was an episode of The Paper Chase (“War of the Wonks”) which featured a computer program that could write well-sourced papers more quickly than humans. But, though that was science fiction when it came out in the 1970s, it appears that that future has arrived, in the form of ChatGPT.

So, what can English language teachers do about it? Or, how can they use it? I honestly do not know: I only started hearing about it a few weeks ago, plus I only recently got a general idea of what it was.

However, I have learned about web search, information evaluation, using Wikipedia, and problem solving. So, based on that, I would advise that teachers do the following about A.I. chatbots like ChatGPT: Decide if they are a problem that they will actually have to deal with. And, if not, don’t worry about them.

    • However, if they are a problem the school might have to deal with, they should pick a couple of faculty or staff members (ideally ones with a good understanding of computer and information technology) to learn as much as they can about them over the next few weeks.
    • In doing so, they would then better understand them, including what potential threats they pose to academic integrity and ways to respond, as well as their potential benefits and how they could use them.
    • Once they had done so, they could share what they’d learned with the rest of the faculty, along with what they thought needed to be done. The faculty could then decide how to implement their ideas.
    • Over the next semester, they could put their plans into effect, while the faculty/staff experts on A.I. chatbots like ChatGPT could also continue learning more about them.
    • At the end of the semester, they could meet, discuss what they’d learned, and modify strategies based on that and new things about it that others had learned.
    • They could repeat this as necessary.
    • They could also talk to other schools, to see what they’d come up with.

By doing so, the school’s staff and faculty would better understand the promise and peril of A.I. chatbots like ChatGPT, while also being able take advantage of the former and mitigating the latter.

The Use of ChatGPT in Creating Multiple Choice Cloze Vocabulary Questions

Judy (Qiao) WANG - Assistant Professor, Center for English Language Education, Waseda University

My colleagues and I have been working on the development of tools for the automated generation of multiple-choice cloze (MCC) questions for vocabulary training and testing in universities. A critical aspect of our project is the identification of source texts that contain target vocabulary words, which we obtain from corpora and the internet.

Upon the emergence of ChatGPT, we saw its potential to support our project and initially attempted to use it to directly generate MCC questions with designated keywords as correct answers, along with accompanying distractors. However, the results were disappointing, as ChatGPT often produced question items that were not in the proper format or context. For instance, there were instances where there was no blank in the question, or the question stem was simply a definition of the word, instead of its actual usage in a sentence. Additionally, the distractors generated by ChatGPT were frequently themselves correct answers when placed into the blank. Despite our attempts to refine our prompts through the use of perplexity or semantic relatedness to restrict distractors, the outcomes remained unsatisfactory.

We are now pursuing a two-step approach to this tool. Firstly, we ask ChatGPT to generate a sentence that includes the target keyword, and then we ask it to select appropriate distractors from a pre-determined list. Our current trial involves experimenting with the academic word list (AWL). If successful, we plan to fine-tune the GPT-3 model, which powers ChatGPT, to better meet our specific needs. Our ultimate aim is to create an app that utilizes either GPT-3 or ChatGPT API, allowing teachers or students to easily generate MCCs for their training and testing purposes.

Memo to Self

Shaun Allen – has taught university and company classes in Tokyo

  1. Review practices in writing assessment:
    What skills and knowledge are being assessed, how, and for what purposes? What constructs are used, what scoring? Are assessment rubrics shared and discussed with the learners? Can they follow the rubrics, applying them in self-assessment and peer assessment? What routines allow learners to track their progress, reflect on their learning, and develop study skills?
  2. Set shorter writing assignments to be done in class:
    Adapting earlier suggestions for evaluated writing being done in class, why not divide major projects such as academic papers into smaller assignments, based on stages in the writing process, to be completed in or out of class, depending on the task? Some serve as short, timed exercises in fluency to be done in class and then redrafted outside. Worksheets set instructions for each stage, with assessment rubrics and other guidance; multiple-choice questions test understanding of the concepts of paraphrasing and summarizing, as well as such cherished concerns as avoiding plagiarism and following proper citation style.
  3. Incorporate chatbots in metacognitive training:
    Remind learners that they, as users of language, are responsible for processing meaning, which machines can’t be since they lack consciousness. Writers who interact with machines therefore need to become attentive readers of fluent and plausible machine-generated text. What challenges might this present to L2 users? What advantages?
  1. Discuss the social risks:
    Be sure to consult with communities of learners and teachers regarding not only the merits of undeniably powerful technologies but also the risks of depending on external platforms the owners of which can change policy at will. Using ChatGPT in the classroom, for example, will require written permission from the institution and consent forms for learners.

How I Deal with ChatGPT in the Language Classroom

Timothy James Byrne - English instructor, UCLouvain, Belgium; he is also a competitive runner

I cannot bury my head in the sand and ignore ChatGPT, much as I also had to get used to dealing with students reverting to DeepL when turning in assignments.

Instead, I raise their awareness, telling them that I am an experienced grader of written work. I know what a student’s output in the other productive skill (speaking) is, and that the transfer between spoken and written output is real and measurable: fluency in the former (often) is proportional to fluency in the latter. Moreover, I always compare work contributed online with work done in the classroom, without outside help. Discrepancies lead to suspicions being raised.

While pointing out the advantages of “AI content writers” (e.g., the availability of templates, their output speed and consistency), I also share that I use the GPTZero detection app. It identifies the levels of “perplexity” and “burstiness” of a text. The former refers to its complexity (the higher the complexity, the more likely it is to have been written by a human). The latter refers to its degree of variation, e.g., between longer and shorter sentences (the more variation, the greater the likelihood it is not generated by a bot).

If GPTZero suggests a student contribution’s low “perplexity” and “burstiness,” I double-check its authenticity by using my own rubric (not shared with my students). If the suspicions raised by GPTZero are confirmed, the student will be accused of plagiarism and the penalty can go from a warning, over a zero, to reporting the student to the institutional authorities.

I have exemplified this procedure below. The starting point is an essay generated by ChatGPT. I processed it through GPTZero and then double-checked the findings using my own rubric. In this case, a red flag shows up and action is required. I then wrote my own essay and processed it through GPTZero too.

ChatGPT generated essay:

Write a 250-word essay in which you develop your stance on the pros and cons of military service

Military service can have both positive and negative impacts on an individual’s life. On one hand, serving in the military can offer numerous benefits such as job training, educational opportunities, and financial stability. For example, many veterans receive financial assistance through the GI Bill, which helps cover the cost of higher education and other expenses. Additionally, military training can provide skills that are transferable to civilian careers, making veterans highly employable.

On the other hand, military service can also have several downsides. For starters, military personnel face the risk of injury or death while serving their country. They are also often subjected to long periods of separation from their families, which can be emotionally and mentally challenging. Furthermore, veterans may struggle to readjust to civilian life after their service is over, and may experience physical or psychological issues such as PTSD.

Additionally, military personnel are often deployed to areas of conflict, which can expose them to traumatic experiences and cause long-term psychological damage. They may also face discrimination and social stigma due to their service, especially if they have served in unpopular wars.

In conclusion, military service has both positive and negative aspects. On one hand, it offers job training, financial stability, and educational opportunities, but on the other hand, it can also have adverse effects such as injury, separation from family, readjustment to civilian life, exposure to trauma, and discrimination. Ultimately, the decision to join the military should be carefully considered, taking into account both the benefits and drawbacks of military service.

Results from GPTZero:

Your text is likely to be written entirely by AI

I double-checked these findings using my own rubric:



ChatGPT generated text

Human written text

Examples from the ChatGPT generated essay above

Presence of idiomatic language








o   There are no idioms and no idiomatic expressions.

Presence of feelings

(empathy, commitment)









o   There is no empathy.

o   The essay doesn’t “communicate” anything.


(lack of insight, understanding, subtlety, inspiration)













o   There is a lack of inspiration.


(comprehensive and efficient without really answering the question)













o   The conclusion is a mechanical repeat of the introduction.

o   There is a lot of repetition throughout.


(nonsensical and incorrect: the “hallucination of fact and fiction”)


Probably not










o   The G.I. Bill only refers to the U.S. This was not asked for in the instructions.

o   The G.I. Bill refused loans to African American G.I.s.

o   The instructions ask the writer to develop their ideas on military service, not on active duty.

o   Conscripts are usually only drafted in times of war (conscripts are not draftees by default).


(e.g., the presence of the personal pronoun “I”)

Doesn’t take a stance

Does take a stance









o   Absence of the personal pronoun “I”

o   The sample essay does not take a stance, while this was stipulated in the instructions.


(formulaic, e.g., too many link words)










o   Too many link words

o   Too many relative clauses (“…, which”)

o   Too much similarity in sentence structure (“… such as …”)

Biased and/or Harmful

(e.g., misogynous and racist)


Not by default








My own essay:

Write a 250-word essay in which you develop your stance on the pros and cons of military service

 In this essay, I will develop my opinion on the advantages and shortcomings of military service.

First, the advantages of conscription. Young people can learn discipline, gain experience and maturity in the military. Some youngsters lack in self-confidence and this can be overcome by doing military service. Physically, one’s level of fitness, perseverance and endurance may increase. There are conscripts who learn professional and soft skills that may prove useful on the workplace. Socially, friendships made during military service may last for a lifetime. Theoretically, social status does not play a role during military service, although there have been examples to the contrary, but these will not be developed here.

Second, the drawbacks of military service. Depending on the unit the conscripts serve in, the hardship of military service can lead to mental problems and physical health issues. Some youngsters have problems conforming to the rules imposed, which they may find meaningless. Moreover, during peacetime and times of global cooperation, military service might not be necessary. In fact, many countries scrapped conscription at the beginning of the 1990s, after the end of the Cold War. Additionally, conscripts might find compulsory military service a waste of time. Alternative forms of national service, such as civil service might be considered more meaningful. Last, armies composed of volunteers, not conscripts, tend to be smaller, more efficient, and more professional.

Taking everything into consideration, I tend to believe that armies should be composed of professional volunteers. While there are advantages to military service, these are outweighed by the drawbacks. More meaningful forms of national service exist, such as civil service, for instance in hospitals and refugee shelters.

Results from GPTZero:

Your text is most likely human written but there are some sentences with low perplexities

ChatGPT: Threat or Opportunity?

Matthew French - currently a lecturer at Hiroshima Bunkyo University

ChatGPT. Even the most techno-illiterate teacher has at least heard of it and the evils this dastardly software is set to unleash upon our classrooms.

After some dabbling, it seems to be the real deal. My initial reaction is ChatGPT will be unlikely to provide much of an avenue for cheating at my current institution because, frankly, the responses are just too natural. So, if I am not immediately concerned with “cheating,” the question then is: “How can I as a teacher embrace the technology and guide my students to use it effectively?” Using technology is one thing. Using it effectively is another.

My first thought was that ChatGPT would be an amazing tool for creative storytelling.

    • Groupwork: Play a “choose your own adventure game”—the teacher gives a prompt; students work together to ask ChatGPT what they should do in response to the prompt. They write a short story or manga about their adventure, referencing their ChatGPT logs.
    • Idea generator: Ask ChatGPT for five story hooks about characters, locations, and MacGuffins provided by the teacher.[1] Come up with a short story outline in your group.

Looking at the somewhat controversial idea of “digital natives” put forward by Prensky (2001), whether you agree with it or not, we are now entering an age where this level of AI interaction will be normalized. Much like fears of student use of the internet and smartphones, students need to be guided in how to harness the obvious power of ChatGPT to produce meaningful and useful results.

[1] A MacGuffin (pl. MacGuffins) is an object or device in a film or book that serves as a trigger for the plot.

If Students Can Write it, They Can Speak about it as Well

Stephan Hughes - English Language and Literature teacher and Adjunct Lecturer in Brazil

Finding ways to curb the Copy-and-Paste culture has been one of our missions as language teachers, especially when students feel they lack the knowledge to put together a well-written text in English. Lecturing them about plagiarism and penalizing those who insist on trying to trick us is an extension of the issue. 

I generally opt for a formative, responsibility-building approach to the issues of plagiarizing work and failing to cite sources. Penalizing students for this kind of behavior is the last line of combat, which starts with raising awareness and discussing consequences for engaging in these practices. ChatGPT is another opportunity to stay on that path. 

ChatGPT has made it inevitable that we English teachers flip and integrate the writing process with a speaking activity. Students can be allowed to research and write their papers with the aid of ChatGPT (after some teacher orientation on responsible use) at home and then present their work orally with the optional help of their notes. They are graded on written and spoken work, which can be checked for originality, if the teacher deems it necessary. 

The quieter students could be given the chance to deliver their presentations privately, at a different time, for example. 

In the end, the process aims to instill in the students a sense of responsibility and accountability for their learning. ChatGPT can be an integral part of this process.

Returning to People

Jim Smiley - Associate Professor at Iwate University researching epistemic cognition

The hype surrounding ChatGPT has been tremendous. Rarely has a piece of technology received so much attention, much of which is at the negative end of the spectrum. Even the blurb sent out to invite teachers to write a piece for this worthy rag notes that ChatGPT is “a threat to ELT writing teachers everywhere.” Another example is seen in Bonnie Stachowiak’s Teaching in Higher Ed podcast (episode 448), which notes that any new technology may be accompanied by two opposite tendencies: those that prohibit and those that promote. My own take on ChatGPT is somewhat more optimistic.

Simply put, we must embrace technology. Not to do so merely instigates an educational arms race. Plagiarism detectors and other prohibitive software must constantly strive to stay ahead of subversive methods. The difficulty of using AI-generation detectors in bilingual educational contexts was highlighted in a recent class. A student submitted a report in English that was far above his usual output. Because I teach content primarily and language secondarily, I accept DeepL translations for classwork whereas I don’t accept plagiarism, a fact that this student seemed to have forgotten. My suspicions piqued, I put the text through two AI content generation detectors. They informed me that the text was 90% fake. Embarrassed and believing that I would reprimand him, the student explained that he had used DeepL to translate his original Japanese. I was more interested in realising that AI content detectors would assess DeepL output as fake. In hindsight, I felt a bit silly when realising that this should be the case as DeepL is a form of AI. The point here is that it’s impossible to detect if the original idea (in Japanese or English) is not fake once DeepL has been utilised.

The positive side of ChatGPT is that I predict that there will need to be much more face-to-face, teacher-learner interaction even in writing classes. Once teachers realise that they cannot rely on AI to detect AI, they will revert to relying on the traditional method: get learners to summarise in real-time in front of the teacher what they had written. If learners cannot do that, then educators can judge the learner’s submission as fake. The old ways are often better but all too often have been bypassed in favour of technological “solutions.” Take, for example, the so-called “on-demand” classes that popped up during the early days of the Covid-19 situation. Many of these reduced learning to faceless and communication-less sentence-level writing and posting. (How much of that was AI generated?). The positive result of ChatGPT ironically may be the need to return to more human interaction during the learning process. For one, I welcome that.

Leave a Reply

Your email address will not be published. Required fields are marked *