So, it has finally happened – the stories that started as fiction in the minds of Asimov and Roddenberry are now becoming mainstream. Our latest guest blog post by our Deputy Head Academic reflects on the challenges and opportunities ahead…
Artificial Intelligence (AI) is infiltrating our lives – and education – in ways that are both exciting and concerning, signalling some seismic, long-term changes. This is a hot topic of discussion amongst our staff and senior leadership team and, just this morning, I led an assembly about AI for all our Senior School pupils. So, what are the key messages we are sharing with our students and our plans as a school community?
Despite being around for quite some time, AI is still very much in its infancy – and it’s already incredibly powerful. Many of our parents work in Artificial Intelligence or its closely related cousin, Data Science, but even those less familiar already interact with AI on a daily basis: voice recognition, Google searches, facial recognition, ‘smart’ home devices, Amazon recommendations, ‘what to watch next’ on Netflix… all these interactions, analysing complex but varying patterns, have their roots in AI. However, much of this work has previously been hidden in the background, masked by ‘Alexa’ or a search bar.
Enter ChatGPT and its cohorts. While the acronym is forgettable (Generative Pre-Trained Transformer) its impact is meteoric. ChatGPT, one of many ‘chat AIs’ currently available for free use, is by far the most well-known. It’s been designed to process text commands and respond in what it considers to be the most useful manner. It is based on data from billions of web pages, which themselves were selected by an earlier AI that finished its work in 2021 – and it learns a little bit more every time it is given a command. As long as the question is entirely text-based, and falls within certain allowable parameters, anyone can ask ChatGPT almost anything…
Chat AIs can understand free written text, including spelling mistakes, slang and idioms, in multiple languages, and use their prodigious databases to respond to that input in an informative and useful manner. This allows people to converse with them, to generate responses that are unique and creative. You can, for example, ask one to write you a poem – or a homework essay. Chat AIs can even create a glossary of terms for some text you have written, offering improvements and synonyms where needed. They can even competently convert text between languages.
Importantly, ChatGPT and its ilk are really just the beginning of what will be a progression of AIs that are accessible, easy to use, and multi-functional. At South Hampstead, we’ve already had staff briefings and are forming a working group to ensure we understand and embrace the potential opportunities for teaching and learning.
When it comes to our students, there are clearly positives and negatives. On the one hand, students now have an incredibly powerful tool at their disposal that can in many ways be beneficial to their learning. When revising they could, for example, use it to generate additional questions for them to work on – and later generate the answers to those questions. It could explain a particular concept, or quickly give them the definition of a key term. It can help them improve a piece of written work and explain the improvements so they can develop in the future. It can also help them come up with ideas for research projects, or even give careers guidance… It is important to note that while none of the current crop of chat AIs have been optimised for education, these will assuredly come in time. To date, the outputs are far from perfect, and will not independently generate work of Grade 9 or A* standard – unless the responses are repeatedly iterated and improved. But the AIs are there nonetheless – and potentially a great source of support and guidance under the right circumstances.
On the other hand, this incredibly powerful tool also has the capacity to do pupils’ work or homework for them – with very little effort. Text generated by ChatGPT is difficult to spot using traditional plagiarism software (although new AIs to spot the work of chat AIs are quickly being developed). ChatGPT can even give long answers to open questions, questions such as ‘write an essay in the style of a 15-year-old student explaining why Henry VIII dissolved the monasteries and the consequences of that action’ – and then be given iterative guidance on how to improve the response until it is at the stage that the user is content with the output. Just take a look at the response to this command, and a requested improvement, to see what can be easily generated using these systems.
Clearly this is a concern, as students using this tool to cut corners are not going to have the same benefits as those who don’t. They will not build their knowledge or revisit what they have already been taught; they will not build skills in, for example, research, essay writing, critical thinking and problem-solving at the same rate as their peers. There is also the risk of students becoming over-reliant on AI technology, as well as the more subtle issue that the AI output will have certain biases, based on its own dataset – a dataset that could easily contain information that is incorrect, prejudicial or inaccurate, despite its creators’ best efforts. It is for these reasons that several schools are blocking the websites for these AIs and banning their use.
As a school, we therefore find ourselves at an interesting crossroads perhaps not encountered since the launch of the world wide web. Once again, we’re witnessing the emergence of a largely untested technology that will undoubtedly play a major role in the lives of our students – with great potential benefits, as well as obvious risks.
At South Hampstead, we have taken the decision to embrace and educate, rather than shun and ignore.
We firmly believe that the best course of action at this time is to allow students to explore this new technology, and for us to work as a community to determine how it can be used to best effect. I have already invited students to meet with me next term to talk about ways in which they have positively used ChatGPT for their learning, so that we can share that good practice throughout the school and start to build a curriculum that includes the prudent and beneficial use of AI.
As a school, we feel that to push students away from these new tools would be detrimental, hindering them as they go into workplaces where artificial intelligences, in all their forms, are increasingly prevalent. We do, however, need to consider and mitigate the risks of this new tech, and continue to educate our pupils about the pitfalls, empowering them to use websites like ChatGPT responsibly, with a healthy degree of scepticism and cross-checking.
We look forward to continuing the work already underway with our staff and students and welcome input from our parent body as well. This is an exciting development with huge scope in every sector. As educators, we want our students to leave South Hampstead as curious, open-minded individuals who are confident in harnessing emerging technologies – productively, creatively, safely and fairly.
Guest blog post written by Mr Westwood, our Deputy Head Academic.