A few weeks ago I visited the BETT trade show in East London with the goal of learning more about how generative AI is being used within the educational sector. Throughout the day I listened to panel discussions, participated in a couple of workshops and went to talk to representatives at a number of company stands. Since the last BETT there has been incredibly rapid development of generative AI and Microsoft and Google along with smaller tech companies are all competing to get at the front of the pack. One insider said he had never seen such a pivot in twenty years and only a lack of experienced, qualified AI engineers was holding them back.
- Generative AI, put simply, is a program that can create totally new output from a vast bank of inputs. ChatGPT is a generative AI program running as a chatbot that many are familiar with - its bank of input is essentially the entire internet and it can analyse and look for common patterns in text and images. Using those patterns it can respond to a user's question or request by predicting which words (or pixels for images) go before and after each word. As you will know if you've used it, ChatGPT gives an impression of intelligent understanding through this process.
- For instance I could prompt: 'Please write a 150 word summary of the British military intervention in Iraq with a reading age of 12 years old' and it will provide me with exactly that based on information it synthesises from the internet.
- I could then prompt it again: 'Please provide another summary of 250 words with a reading age of 15, and one final summary of 500 words with a reading age of 18'.
- You can see in this example - which I actually did recently - there is great potential for preparing differentiated lesson materials to provide greater access to learning (it goes without saying that a specialist teacher has to review outputs for balance and judge if they are right for their students).
- I'm not alone in using generative AI in this specific way - Microsoft Education is a similar platform to Google Classrooms but with an integrated generative AI program (Microsoft Copilot) so teachers can create bespoke texts for students at different reading ages. The AI part of the platform is still in its infancy but in a year things will likely look very different.
I attended several panel discussions and the general tone of debate around AI was, as expected for the context, boosterish. However, many ideas seemed grounded in reality and responsive to educational problems: Could AI be used as a personal tutor for school avoiders to help them keep up whilst they worked towards rejoining mainstream schooling? Could AI speed up some administration tasks to free up overburdened teachers to do more teaching? Could AI provide hyper-individualised learning that accelerates progress by engaging students more in and out the classroom?
As well as Microsoft and Google I spoke to a representative of Redmenta, a Polish company which won the 'AI in Education' award at BETT 2024. From what I could see the platform did everything that a teacher could do, creating straightforward reading comprehension tasks and worksheets but in literally seconds rather than hours of a teacher's time (likely outside of work hours). In this situation the teacher becomes an expert editor of material rather than a copywriter. The potential for AI to increase productivity across industries is very clear but in education where recruitment and retention is the number one employment-side issue, this is even more relevant. For this reason, one speaker at a talk hailed it as a 'saviour of education'.
Other ideas from panellists were more leftfield and less convincing: one speaker suggested that the secondary stage could become more like primary with one teacher per class and AI doing the job of specialist as they changed subjects across the day.
Later in the day I participated in a Microsoft workshop to have a go with Copilot. It runs on an enhanced version of OpenAI's ChatGPT and also provides references and web links for the responses it gives. I can see this becoming the standard type of web search tool for internet users in the future - in contrast to the normal format of paginated search results. As with many of the demonstrations I saw, presenters were vague on the details of how it could be used in teaching but big on overall vision.
One of the key takeaways from my visit came from something that was said at a Google for Education panel discussion with several practicing teachers and leaders - I'm paraphrasing but a teacher said 'whatever technology you are putting into your school you need to make sure you are identifying an area of Teaching and Learning to improve first and then finding the right tech to help improve it - not just shoving in the latest tech for the sake of it'.
A digital strategy in a school cannot sit separately to its teaching and learning strategy - it must be fully integrated just like any other pedagogical idea the school wants to introduce. In practice that means training staff to use the tech properly, then having patience to stick with it and develop it so that every teacher is using it effectively and consistently, then attempting to measure its impact on students progress and reflecting deeply before moving forwards.
One of the panellists earlier in the day made a prediction that in one years time there will be a greater divide between schools that embrace it and those that don't, but in ten years time everyone will have adopted it. I think this is an easy prediction to make given AI is likely something we will have to reckon with in all areas of work and life very soon. Whatever your current thoughts on AI in education, it is unlikely to be a passing fad and a forward thinking school will be seeking ways it can help students achieve success right now.
Thank you,
Ben
No comments:
Post a Comment