ChatGPT. If you haven’t heard of it yet, it is time to come out from under that rock. Seriously, come on, come and join us… it isn’t that scary!
It seems that every mainstream and social media service over the past month or so has been talking about ChatGPT and how disruptive it is going to be to education as we know it. Education departments in QLD, NSW and VIC have already banned it – so it must be really dodgy and dangerous!
So, what is ChatGPT?
ChatGPT is a type of Artificial Intelligence – developed by OpenAI who is an “AI research and deployment company”
OpenAI’s mission is to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity.
We will attempt to directly build safe and beneficial AGI, but will also consider our mission fulfilled if our work aids others to achieve this outcome.
https://openai.com
The current attraction of ChatGPT as a type of Artificial Intelligence is that (as the name suggests) it works in a conversational manner. You ask a question in normal conversational language which is interpreted by ChatGPT to provide you with a response, making it extremely accessible to most people to use. I wasn’t sure what GPT meant, so, I asked ChatGPT:
Dave: What does the GPT in ChatGPT stand for?
ChatGPT: GPT stands for Generative Pretrained Transformer
Ok great… that’s cleared that up! Not quite…
Dave: And what does that mean?
ChatGPT: GPT refers to a type of artificial intelligence language model developed by OpenAI. The “generative” part means that it can generate new text based on the patterns it learned from the large amounts of text data it was trained on. The “pretrained” part means that it was initially trained on a massive corpus of text data before being fine-tuned for specific tasks, such as answering questions or generating text. The “transformer” part refers to the architecture used in the model, which is based on the transformer network introduced in the paper “Attention is All You Need”.
Thanks, that’s much better! So ChatGPT can provide responses based on a huge cache of information, learns from the information it has access to, and can provide those responses in the form of a coherent answer, rather than the user needing to interpret a range of search results for themselves like you do when you “google” something.
The transformer architecture is different to a normal neural network which will process data sequentially – i.e. each new piece of data adds to the understanding built before. The processing is done by a transformer network in parallel – which means that all of the data is considered at the same time… allowing ChatGPT “to understand the relationships between different parts of the input, leading to better performance”, particularly when answering natural language questions. (Thanks again ChatGPT)
And so, if you’re following along you’re probably starting to realise what some of the fuss is about from an education perspective, and why administrators have got themselves in a bit of a flap. Here’s a new piece of Artifical Intelligence (AI) which students could potentially use to answer questions provided by teachers. It can provide simple answers:
Analyse more complex questions:
Refine answers:
And provide complex, essay-like responses
It is pretty easy to see from these very simple examples that ChatGPT is a tool which could be used by students to “cheat”. But is blocking the answer? Should we be worried and concerned? Let’s go a little deeper and think about what this might mean for education.
As a starting point it is helpful to place AI in the context of a Christian response to digital technology. In the Genesis story of creation, humankind is given responsibility for cultivating the earth – unpacking the latent potential that exists. This includes using metals to develop tools, making bricks etc for shelter, and extends to the potential encompassed by digital technologies. We know, however, that because of the fall of man, our unpacking of this latent potential is often done in ways which are not aligned to God’s will for His creation. Derek Schuurman unpacks this far more effectively and eloquently on the Christian Scholars website – Technology and the Opening Chapters of Genesis.
So AI – and ChatGPT is no surprise to God. However, we must tread with care. I frequently have conversations with teachers and school leaders about maintaining a distinctly Christian response to digital technologies and have lost count of the times I’ve been met with a response that goes something like:
“But Dave, technology is just a tool – it is value neutral… it can be used well, or used poorly, but that’s entirely in the hands of the user”.
While it is true that technology can be used well, or used poorly it is important to understand that technology is not neutral at all. It is value laden. It has a significant shaping effect on the users as each technology is developed with a particular purpose, goal and even worldview shaping its design. This focus then shapes the way the user interacts with the technology.
Perhaps this is more easily identified in AI systems where the direction and output of the system is obviously limited by the data contained in the system. The designers of the system need to make initial value judgements on behalf of the system, effectively teaching it how to “think”. ChatGPT itself is “trained to decline inappropriate requests”. The creators have embedded values into the system, but who has decided what is inappropriate and what is not?
As adults it is important that we recognise these inherent values embedded into AI and discern the impact of these values on the output produced by the system. We shouldn’t be afraid of the system, nor should we embrace the system without due consideration. Instead, careful analysis and critique can help us as Christians use AI like ChatGPT in a way which gives honour to the Creator.
And so that gets us to the education part. What are our options, and how might our schools respond to this changing landscape?
Should we block it?!
I feel like we’ve been here before. It seems that when something new enters the education landscape the initial response of many is to ban.
Here’s a new technology which can do work for a student! We should ban it! Straight to the IT department – let’s block it!
We can see a history of this in education, and it hasn’t (ever?) ended well for the “blockers”. Many technologies we take for granted now were once resisted. Ball point pens, chalkboards, whiteboards, interactive panels, calculators, computers, laptops, iPads/tablets. Each of them has been new at some point, been resisted as a new innovation, and subsequently (with all of their included benefits and flaws) become part of the accepted landscape.
Artificial Intelligence such as ChatGPT might be a bigger revolution to education than any of those previous, but I don’t believe blocking it is the answer. While ChatGPT might be the first to make headlines, it most certainly won’t be the last. Will the blockers block the next, and the next? What about the one after that?
Our response ought to be more considered than this response, which some might consider lazy. Our responsibility to students is always to teach. To teach them about this new technology, and to teach them how to use it responsibly as Christians.
What does that look like?
This indeed is the tough question we need to ask ourselves. Here are a few ideas I have which might start us on this journey…
1. Understand ChatGPT
Ok, so this might sound obvious, but I think it needs to be said. Teachers need to understand ChatGPT. This doesn’t mean we leave them alone and send them off to “check it out”. We need to be deliberate in our support of teachers, training them how to use it well, what the limitations are, outlining the shaping effect of the technology and demonstrating its amazing capability.
2. Use ChatGPT
Teachers (and schools as a whole) should use ChatGPT. How? There are already many ideas others have suggested including:
- Using ChatGPT to modify existing assessment / tasks for students with particular special needs
- Writing policy documents
- Developing marketing content
- Modifying email communication to make the tone more/less formal
3. Teach (with) ChatGPT
As our understanding of the technology improves, so we should consider how it can become a part of our teacher toolkit:
- Teach students how to use (and refine) the prompt effectively to gain the desired output across all KLAs
- Compare the responses given to poor prompts with the responses to well crafted prompts
- Have students develop a response with ChatGPT and then refine that response to more effectively answer the required question. Analyse strengths and weaknesses of the original response.
- Provide students with a ChatGPT response and reverse engineer the prompt used.
- Students to analyse a ChatGPT response for bias, examine and explain the worldview inherent in the AI
- Have students write their own response in class, then have ChatGPT refine the response. Analyse the difference.
- There are plenty of other ideas out there – here’s a great starting point from Leon Furze: https://leonfurze.com/2023/01/23/practical-strategies-for-chatgpt-in-education/
4. Review Assessment
Clearly, there are some potential challenges for assessment in schools, particularly with the traditional “take home” types of tasks which might be particularly prevalent in the humanities. (Incidentally, ChatGPT has far greater capabilities than this – it can create code in a variety of different languages, for example.) Without opening the can of worms that is assessment in schools, teachers will need to carefully consider tasks given and test themselves whether ChatGPT will be able to provide responses on behalf of students. Consideration will need to be given to both formative and summative assessment to effectively support student learning.
This article is already much longer than I had intended, but there’s so much more. ChatGPT has just (at the time of writing) released a subscription model for a premium service promising faster response times and priority access. Google is launching a competitor called Bard in the coming weeks. The space is constantly changing.
My final encouragement to Christian Educators is to take a look. Please resist the temptation to try and block this technology out of fear. Instead, engage and learn. Consider what value it might bring to your classroom or institution, but do so in a thoughtful and discerning way.
Did you know it took ChatGPT only 5 days to reach 1 million users? Facebook took 10 months…
Thanks Dave, that is really concise and thoughtful.
Hi Dave,
This is a great blog! I love the easy way you’ve explained it and particularly how you have helped to point our hearts in the right direction, as Christians.
We have been playing with ChatGPT at Wellington Christian School as a staff. As a K-6 school we realise that this will be the norm for our kids when they enter high school. I initially introduced it to the staff, none of whom had heard of it, with a devotion. I asked the question- Does all creation shout His name? in reference to IT and other technologies. I asked ChatGPT Who is Jesus Christ? The resulting conversation was very cool. The fear is gone 🙂
I have a question about copyright in using ChatGPT. We experimented with writing a unit of work for English and linking it to the threads from Transformation by Design. The initial responses have been mind bogglingly accurate. We also asked for activities that are written for a Christian perspective. Also incredible! I’m wondering if there are problems or copyright issues when the program uses readily available information from published works like TBD??
The capacity to use it for programming is amazing!
Jo Blatch
Principal
Wellington Christian School
Hi Jo,
Thanks for reading! I’m glad you have found the article useful and am excited that you’ve started the conversation around this at Wellington.
I think you’ve raised a really good and tricky point with relation to copyright, after all, we want to teach our students to effectively and accurately attribute the work of others. The challenge with ChatGPT (and other AI) is that most of the time we don’t know what the true source of the response is. Sure, we can attribute ChatGPT, but the AI has just brought the ideas of others together for us (often in a very useful and coherent way). How do we acknowledge the original author of those works? I don’t know that we can. If someone has a better answer I’d love to hear!
In the course of my work I asked ChatGPT to provide me with a specific framework recently – which it did very usefully. I suspect it probably came from a single author elsewhere – but I don’t know.
Definitely something to continue wrestling with as we ride this rollercoaster!
Dave