Technology companies have a preoccupation with “time-saving”, and recent reports have centred on the number of hours per week that teachers might claw back by using AI.
In a recent article, I questioned whether “time-saved” is the best measurement when dealing with teacher workload. I think that there are ways GenAI could improve the quality of teachers’ day-to-day experiences, but we need to look for areas beyond the low-hanging fruit of lesson plans and admin documents.
Read more from previous articles on AI and teacher workload:
Based on research into teacher attrition and burnout, I suggested six areas where I think we could measure and improve outcomes for teachers:
- Achievable, well-resourced work
- Genuine professional autonomy
- Meaningful and purpose-driven work
- Targeted feedback and growth
- Support and collegiality
- Trust-based accountability models
Over the past three years I’ve worked with dozens of schools, and have supported many pilot studies related to AI and teacher workload. The following examples are drawn from those experiences of how educators are using GenAI to make meaningful improvements beyond “efficiency”.
Interested in how AI is being used to manipulate the media and spread misinformation online? Make sure to grab the free 20+ page resource How to Spot a Deepfake by signing up here:
Well-resourced work
One of the biggest challenges facing schools right now is rising attrition coupled with falling numbers of new teachers entering the profession. Rates of attrition within the first few years of teaching are also high compared to over industries, which means that even if a school does secure new teachers, the turnover can be unsustainably high.
One factor in early career teacher retention is providing the space and mentoring to support new teachers in developing or accessing existing resources. However, many schools do not have the capacity or human resources needed to implement effective mentoring programs.
We might bridge some of that gap by pairing a faculty mentor – even one with limited time available – with GenAI in order to collate and identify teaching resources.
For example, the various “Deep Research” products from OpenAI, Perplexity, Google and Anthropic can be useful for scouring existing curriculum websites to identify suitable resources that have already been audited by other educators or authorities. These resources are often buried behind pages and pages of drop down menus and nested links on websites like the ACARA v9 Curriculum pages. AI can help unearth them.
The Deep Research models work well for this task because they are thorough and present the information in a well-structured and accessible fashion. The final output can be exported as a PDF (or via Google Docs for Gemini) and saved as part of the faculty’s documentation for future reference.
Professional autonomy
In numerous studies, teachers report a greater sense of self-efficacy or general wellbeing when they have increased agency and professional autonomy, including in choices like how and when they conduct their professional learning.
Unfortunately, professional learning is very prescribed in some jurisdictions, with training being limited to “authorised” programs and delivered in bulk whether teachers really need it or not.
Rather than one-size-fits-some approaches to PD, GenAI can be used to allow more individualised professional learning programs. Staff could be encouraged to engage with AI as part of their annual performance review cycles to identify and support PD programs.
This example is an adaptation of a custom GPT being trialled in a school to help with those processes:
Meaningful work
Wellbeing is tied to a sense of purpose, and in education that frequently comes from a passion for a particular subject, topic, or discipline. I doubt many teachers enter the profession because of a desire to complete paperwork, but there are plenty who start teaching because they are passionate about reading, writing, mathematics, sciences, the arts, and so on.
Allowing staff to play and experiment with GenAI in the context of their disciplinary expertise can lead to great things: engaging materials, fun side projects, and uses of AI that a non-expert would never have thought of. I see this time and time again when teachers outside of my own disciplines surprise me with how they’re using the technology.
This example shows what happens when a science teacher plays around with a more powerful model like Claude 4, for instance:
In this example a (hypothetical) Physics teacher uses Claude’s “artifact” feature to create an interactive in order to help explain a concept from the Year 12 exam. This would only be possible with the combination of domain, situated, and technological expertise needed to verify the accuracy and utility of the finished product.
Targeted feedback and growth
Mentoring and coaching experiences for teachers – when they’re used well – often include opportunities for reflection on challenging moments such as behaviour management, dealing with parental concerns, or working with students with challenging additional needs.
Many AI platforms now feature an advanced speech-to-speech voice mode which can be used to simulate some of these issues. Increasingly, these voice modes can be paired with the other features of AI such as document upload or internet access to provide additional context to the model during use. These simulations could be used alongside those peer mentoring experiences, or even just in a conversation with a faculty leader, assistant principal, or colleague.
This example shows advanced voice mode in ChatGPT used to simulate a call from a parent who is unhappy about their child’s support in a class, with me role playing the faculty leader responding to the call. During the scenario, I ask the model to pause and browse the internet to get some specific context for the conversation.
Following the conversation, the voice mode window can be closed and the transcript could also be used as part of those ongoing mentoring and review processes.
Support and collegiality
As well as individualised professional learning programs that allow teachers to specialise, there is a lot of power in collective PL models including professional learning communities (PLCs) and Communities of Inquiry/Practice (COIs/COPs).
GenAI can definitely be useful in helping to manage the administrative elements of shared practice, and also to extend the materials and resources from a COI or similar process into something even more useful.
Many people are familiar by now with Google NotebookLM and the “generate a podcast” feature, but it also makes a useful tool for managing a collaborative inquiry, PLC, or action research project amongst a group of staff.
Trust-based accountability
This last one is a little out-there, and probably not a “right now” use of the technology, but I want to demonstrate something that I think will absolutely be possible in the near future.
There is great potential in using lesson observations and reviews in the context of mentoring and coaching conversations, including lesson recordings. Unfortunately, there are lots of barriers. First, schools have to have processes in place to ensure that mentors are free when mentees are teaching. Then, the mentor and mentee have to have the time and capacity to review the video together, and a framework through which the lesson can be evaluated.
Increasingly, GenAI models are multimodal and can take text, image, audio, and video as input. Theoretically, these multimodal models could be used to assist in the process of gathering and organising recorded lesson observations. There are of course important caveats, including privacy, data handling, and the purpose of the recording. Recorded lesson observations are not a form of surveillance, and are not necessarily tied to performance reviews. Any platform used to record or analyse footage would need strict privacy controls, and no third party sharing of data. And consent would be required from anyone included in the footage.
Still, the recording of lessons for observation and feedback is not a new idea, and it could be extended with a suitable multimodal GenAI model. The example below demonstrates Google Gemini analysing a YouTube video of a recorded lesson as an example of what this might look like in the future.
Beyond time-saving
Many of the examples above will take time to set up, and perhaps even require additional time to use to the fullest. But “saving time” is not the point: it is the quality of the time we do spend.
Where each of these examples make a real difference is in the coordination of GenAI with human mentors, coaches, and colleagues; for example, supporting early carer teachers by allowing faculty leaders to have more meaningful conversations or using AI to identify or create resources that support purposeful work.
Time-saving seems like low-hanging fruit. It is measurable, and makes for great advertisements for technology. But it doesn’t actually shift the dial on reducing burnout or making teachers happier and more inclined to stay in the profession. For that, we need to find ways to pair the technology with the expertise we already have in schools.
I’d love to hear any more examples you have of ways teachers are using GenAI beyond time-saving.
Want to learn more about GenAI professional development and advisory services, or just have questions or comments? Get in touch:
Subscribe to the mailing list for updates, resources, and offers
As internet search gets consumed by AI, it’s more important than ever for audiences to directly subscribe to authors. Mailing list subscribers get a weekly digest of the articles and resources on this blog, plus early access and discounts to online courses and materials. Unsubscribe any time.
Leave a Reply