Leon Furze

This is the fifth and final post in the series Teaching Writing in the Age of AI. Check out the other posts in the series:

• Teaching Writing in the Age of AI
• Talking about academic integrity
• Asking questions
• Assessment and “cheating”

In this series, I’ve explored how teaching writing might change as the adoption of Artificial Intelligence accelerates. Late 2022 was a watershed moment in AI, with ChatGPT pushing generative AI into the mainstream. But there have been many technologies other than OpenAI’s chatbot released in the past few years, and plenty of them will have an impact in education.

As an English teacher, I’m particularly concerned with the impact these technologies will have on how we teach writing. I think that English teachers are well-placed to adopt and even develop these technologies. To some extent, I think we will have to learn to adopt AI. On the other hand, there will always be room for traditional methods of writing instruction which are not reliant on technology.

Teaching writing has become, in many education settings, primarily for knowledge assessment. In this final post, I’ll explore how AI has changed that, and why it may well be a very good thing.

The Calm Before the Storm

As a system, education is often accused of being slow-moving, too traditional, and reliant on industrial-era methods and practices. In some ways, I agree. Standardisation, high-stakes testing, and accountability have made many aspects of teaching simultaneously complex and dull. For example, I had the luxury of working in a school that wasn’t too focused on NAPLAN results – the Australian standardised test that runs in Years 3, 5, 7, and 9 – but many are not as fortunate.

We’ve seen recently how testing like NAPLAN can drive the narrative around teaching, with media reports ranking schools and publishing “league tables”. It’s an agenda that particularly impacts English teachers: NAPLAN literacy outcomes tends to dominate the discourse around reading and writing. In many schools, teaching writing – at least up to grade 9 – means teaching the persuasive and narrative styles required by the test. It’s demoralising for English teachers, and despite repeated entreaties to not “teach to the test”, it’s inevitable that schools aiming for high ranks will require teachers to do just that.

On the other hand, in my practical experience in the classroom, I’ve seen amazing and inspiring lessons come from English teachers that defy standardisation and promote creativity and critical thinking. These pockets of excellence are clearly visible at events like the VATE and AATE state and national conferences, and the recent VATE VCE Days. These are teachers who have not succumbed to the pressure of NAPLAN and senior exams, and have been able to work within the supposedly rigid, outdated system.

Teachers have always had to navigate this balance of producing creative, engaging lessons with standardised curriculum and testing requirements. This status quo, however, is about to be disrupted by technologies which render the kind of writing assessed by NAPLAN obsolete.

Into the Vortex

It’s already clichéd to talk about how disruptive ChatGPT has been. Articles beginning, “In late November 2022, tech startup and former non-profit OpenAI launched ChatGPT onto an unsuspecting world,” were published everywhere from Wired to the abc, TechCrunch to the Sydney Morning Herald. The fact that ChatGPT garnered so much attention in both the technology sector and the mainstream media is what differentiates it from previous technologies.

But many educators have had their eyes on this tech for a while now. As early as the start of February 2021, my supervisor Lucinda McKnight published this article in The Conversation where she claimed “machines are also learning to write, so effectively that within a few years they may write better than humans.” Looking back, “a few years” was an underestimate. Less than two years after the article was published, ChatGPT took the existing technology, slapped a user-friendly interface on top of it, and threw it out into the wild.

You can argue that ChatGPT doesn’t write that well. You can point out that, technically, it “can’t write essays“. But there’s no arguing that the chatbot can produce writing which is of a higher quality than many people can produce. It is consistent, grammatically correct, and can adapt many styles. The technology also “hallucinates”, produces biased output, and isn’t particularly creative. That doesn’t mean that it can’t write, though. It just means that, at this very early stage in its development, it isn’t perfect.

The release of ChatGPT has sent shockwaves throughout the educational community, with many educators scrambling to adapt to this new technology. That swirling vortex has suddenly opened up right there in our classrooms, and we’re struggling to keep our footing. While some students may have been using GPT-based technologies in the classroom before ChatGPT’s release, its high profile launch has made AI much more visible and accessible. AI is no longer some distant, sci-fi concept – it’s here in our classrooms, whether we invited it or not.

The Eye of the Storm

Right now, we’re in the middle of a hugely disruptive period when it comes to AI in the classroom. Many of our traditional methods of assessment and instruction won’t work in the face of a technology that can easily produce student-level writing with a few taps on the keyboard.

That doesn’t mean the end of High School English, in spite of claims to the contrary in the media. In fact, it can and should be treated as an opportunity. Like I mentioned earlier, English teachers are well primed not only to ride this storm out, but to benefit from AI technologies. Here are a few reasons why English teachers are well-placed to deal with generative artificial intelligence:

• English teachers are trained to read and analyse texts critically. This skill is more important than ever in an age where AI-generated texts can easily spread misinformation and bias.
• English teachers are also trained to teach persuasive and creative writing, skills that are difficult for current AI technologies to replicate without careful prompting. By emphasising these skills in the classroom, we can help our students develop their own unique voices and perspectives.
• English teachers (and their school library colleagues) are adept at teaching research skills, which are essential for navigating the vast amounts of information available online. By teaching our students how to identify credible sources and fact-check information, we can help them to mitigate some of the limitations of current AI like Microsoft’s new Bing.
• English teachers have always taught ethics. AI is a highly problematic technology, with ethical concerns that run deep. From the well-publicised bias to lesser-known impacts on human labour and the environment, AI ethics needs to be thoroughly explored in the classroom.

Add to this list the fact that generative AI is entirely built on language, and you have a clear rationale for English teachers being among the first to adopt the new technology.

When the Dust Settles

Once the hype cycle of AI is over, and the dust has settled, we’ll be able to get a clearer picture of how, when, and why to incorporate AI into the classroom. We’ll also have an indication of when not to include the technology.

It seems clear already that students will be using generative AI in all sorts of applications, including text and image generation. Many universities have already adopted policies which permit students to use AI as long as it is correctly acknowledged. Workplaces from marketing agencies to law firms have started to explore the technology, and big companies like Coca-Cola are on board.

Right now, there is still a lot of speculation about what education looks like post-ChatGPT. I have seen many discussions about academic integrity, the need for updated assessment methods, and a shift away from content-based instruction and towards skills that hone students’ critical and creative capacities.

There is also vast speculation around the technology itself. Remember, ChatGPT was only released a few months ago. In that time, we’ve seen announcements from Google that its chatbot Bard is on the way. Microsoft has released New Bing into the world, and indicated it will integrate GPT into its Office apps like Word and PowerPoint. Meta has released a powerful language model for research purposes. And OpenAI has been building towards the release of GPT-4, which will supposedly be an order of magnitude more powerful than the current model which underpins ChatGPT.

Let a few weeks or months go by, and the AI landscape will look very different to right now, at the end of February 2023. If you’re feeling overwhelmed by the storm of information and excitement (and fear) about AI, you’re not alone. But this will pass, the dust will settle, and these technologies will become as commonplace as word processors and internet search engines.

Writing the Future

While we wait for the hype cycle to die down, there are a few practical things that can be done right now. Again, I’m pointing this advice at my most familiar audience: English teachers. But I think that much of this is relevant for educators of all stripes.

• Do a little professional learning: At this early stage in the technology’s lifecycle, a little PL will go a long way. I’ve been running sessions on Practical Strategies for ChatGPT, as well as sessions in schools. There’s also lots of great, accessible, and free professional learning out there, including Nick Jackson’s course for teachers.
• Review your assessment practices: Look at how you assess writing. Are tasks designed in a way that can easily be completed by generative AI? If so, what could you change? My previous post in this series had lots of practical advice on how to adjust tasks.
• Don’t revert to exams: Long before AI was a thing, I posted a popular article suggesting that conducting school assessed coursework (SACs) under exam conditions is not the best approach. It’s a contentious point, as most schools run SACs as mini-exams. Even in the face of AI, supervised, timed exams are not the answer. These kinds of assessments disadvantage many students – find alternatives.
• Chuck out TEEL: I’ve written plenty before about my opinions on formulaic structures like TEEL. I won’t keep harping on about it.
• Foster creativity: AI technologies can often produce formulaic writing that lacks the creativity and originality that good writing requires. By emphasising the importance of creativity in the writing process, we can help our students develop their own unique styles and perspectives that are difficult for AI to replicate.
• Teach digital literacy: As AI technologies become more prevalent, it’s increasingly important for students to develop critical digital literacy skills. By teaching our students how to evaluate online sources, identify bias and misinformation, and fact-check information, we can help them become responsible and informed digital citizens.
• Emphasise the importance of ethics: AI technologies have the potential to reinforce bias and perpetuate inequality if not designed and implemented ethically, and that’s not even the greatest ethical concern. By teaching our students about the ethical implications of AI in the classroom and beyond, we can help them become responsible and thoughtful users of technology.
• Collaborate with AI technologies: Rather than seeing AI as a threat, we can choose to collaborate with it to enhance our teaching and learning practices. There are already many great examples out there for how teachers can incorporate AI into the classroom.

The future of writing, as far as I’m concerned, looks bright. I’m not threatened – as an educator or an author – by machines replacing essential, human skills. Writing is not just about demonstrating knowledge. Teaching writing in the age of AI will definitely change. If English teachers take the reins on these technologies, then it will change for the better.

Got questions or comments about this article? Want to get in touch to discuss teaching writing or anything else about AI? Get in touch using the form below: