Dr Leon Furze

Earlier this week, I attended a conference on AI run by The Hatchery in Melbourne alongside some great presenters and panelists representing a cross-section of K-12 and higher education. During one of the panels, Emily Ashcroft of University of Canberra mentioned that many of the voices we were hearing came from the Independent sector, and there seemed to be a limited amount of feedback from low SES schools or schools in regional areas.

It’s certainly true that Independent schools are overrepresented in these kinds of conferences right now, and I think there are many reasons for that. For a start, the independent school sector across Australia never banned ChatGPT in the early days, whereas most state departments of education did. Having those extra few months to explore the usage of tools means that in these conferences, where educators are encouraged to share their experiences in the classroom, the majority will of course be Independent and Catholic.

It’s a simplification to say that Independent schools are wealthier, since the sector includes many small, regional, and secular or religious schools which don’t fit into the stereotype of a “rich private school”. But there are of course Independent schools that have access to the resources to also experiment at the higher end of the AI spectrum, such as building their own custom chatbots and working closely with organisations like Microsoft to test sophisticated or costly systems. We’ve heard a lot from those schools in recent months, but the question remains: where are the other voices?

It was a great question, and I couldn’t help but jump in and comment on one sector that hasn’t really had much coverage in these conversations about AI in K-12 education: Vocational Education and Training (VET).

Vocational Education and Training

VET is Australia’s certificate system for trade-based qualifications, but it extends beyond the more commonly thought of trades such as building and construction, electrical engineering, and automotive, and encompasses a wide range of subject areas and disciplines. Towards the end of my teaching, I got the TAE Certificate IV required to deliver VET courses, and taught music industry. There are VET courses in business, finance, ICT, engineering, healthcare, and education – a huge range of domains.

Since beginning my consulting work at generative AI, I’ve worked with a number of vocational education organisations, including most recently running a series of webinars for the VET Development Centre in Melbourne on areas like generative AI assessments in vocational education, and with one of my local regional clients, Rural Industries Skills Training (RIST), an agricultural studies provider. I’ve recently had a great conversation with VET hospitality teachers at one of my client schools about how these technologies might be used as part of courses like kitchen operations.

Why Does Generative Artificial Intelligence Fit in Vocational Education?

Whether true or not, vocational education is often seen as an alternative to mainstream education, and VET courses often attract students who are not aligned with VCE or HSC pathways into higher education. That’s particularly true of many of the trade-based courses, although it is worth pointing out that all VET courses, from building and construction through to the music industry, are perfectly valid pathways into many higher education institutes and are respected qualifications both for young learners and in adult education.

When it comes to GenAI, much of the conversation in education has been focused on academic achievement, perceived threats to academic integrity, and the risk that this technology poses to written assessments. I think that vocational education actually offers some fantastic alternative forms of assessment which are less vulnerable to generative artificial intelligence. If you’re not familiar with vocational education, assessments are often incredibly rigorous, sometimes to the point where the paperwork on evaluation and assessment is significantly longer than the assessment itself…

The forms of assessment are very interesting and largely geared towards practical application. Here are a few examples:

• Observation checklists
• Role plays
• Scenarios
• Workplace activities
• Reports from employers

Vocational training, by nature, is practical and geared around skills which are needed for the particular job role or discipline being studied. Mainstream education, by contrast, is focused predominately on subjects and content. National and state curricula are carved up into curriculum areas familiar to anyone who’s been through the education system: English, science, humanities, arts, maths, technology, languages, and religion.

The division of curriculum areas into subjects leads inevitably to curricula which focus on discipline-specific content. While there are, of course, skills unique to each discipline, and the Australian Curriculum is deliberately moving away from content-heavy curricula, it’s very hard to disentangle content from a subject-based curriculum. This, in many ways, is why artificial intelligence has been seen as a threat. Generative artificial intelligence systems like GPT are predicated on the fact that a machine learning model can consume an enormous amount of content, learn the rules and connections of that content, and then spit it out on demand. Trained on a corpus that includes almost the entire internet, large language model technologies like GPT can generate more content than any individual teacher.

The Rise of the Chatbots

One side effect of this has been a focus in educational contexts on creating educational chatbots. I’ve written about chatbots before a few times, calling them both a dead end and questioning education departments’ attempts to build safe chatbots rather than focusing their attention on educating about the risks of artificial intelligence.

But I’m just an old man yelling at the clouds, and while I fold my arms and grumble about the tedious, transactional kind of learning that comes out of chatbots; their tendency (despite everybody’s best efforts) to hallucinate; and the frankly laughable oxymoron of an “unbiased AI”, the rest of the world is getting on with the business of deploying chatbots in as many educational contexts as possible.

Except, when I talked to vocational teachers and students, they couldn’t care less about chatbots. Where are all the chatbots in vocational education and training? Vocational education and training is not a knowledge-heavy, curriculum-based subject. It is a vocation and skills-based subject area, and students, with a few exceptions, are not required to rote learn or memorise large passages of information (although there are strict standards around minimum levels of literacy and numeracy). Assessments are not contingent on lengthy written essays.

When I talked to vocational teachers about artificial intelligence, there’s only really one question: How can these technologies help my students to get the job done?

GenAI in Vocational Courses

The agricultural studies provider I work with is very interested in exploring how generative AI technologies can assist students with low literacy in communicating clearly with employers, suppliers, and customers. They’re interested in exploring how generative AI voice transcription can be used as an assistive technology for students who can’t or won’t put pen to paper. Although a written answer is required by the assessment mode, these students won’t sit at a computer and laboriously type in a paragraph on the various merits of different fertilisers for soil quality.

But they may be able to perfectly explain those qualities whilst out in the paddock standing in front of a truck full of fertiliser, capture that explanation in an audio recording, transcribe it, submit, and move on with the job.

In kitchen operations, can artificial intelligence be used to help coach students in getting better use of existing technologies like spreadsheets and databases for managing aspects of the business? Yeah, absolutely.

In music industry, can artificial intelligence tools incorporated in music software lower barriers to entry for complex applications which can be used to create electronic music? Already happening.

In arts and design-based courses such as the Certificate II in Creative Industries or the Certificate III in Screen and Media, can artificial intelligence be incorporated in the industry-standard platforms which allow students to create images, audio, and video of impressive quality in the short time allotted to them in the classroom during the VET program? Yep. The technologies already exist. We just have to allow the students to use them.

Only once has the conversation about chatbots come up with VET, and that conversation went something like this:

We have reams of curriculum materials aligned to the various certificate studies that the students are expected to just plough through. But some of these students struggle with reading, some couldn’t care less about wading through endless PowerPoint slides, and some have just got better things to do – they’re already working full-time jobs. Can chatbots help to just pull out the specific information that a student needs from these curriculum documents at the exact points of time that they need it most, wherever they are, whether they’re in the classroom or out in the paddock?

I think this is a fantastic way to think about generative artificial intelligence technology, even chatbots. Using AI, or framing AI, as a point-of-need assistive technology and then teaching students how to use it if and when they need it is much better, in my opinion, than the idea of having rows upon rows of students staring at a laptop, individually interacting with chatbots to learn curriculum content which two years ago, or 20 years ago, or 100 years ago, might have looked exactly the same.

The Need for Diverse Perspectives

We absolutely need to hear more from diverse education contexts and different groups of students in conversations about AI. In Australia, the Independent sector may have had the jump on developing AI systems, but the departments of education have now relaxed the ban on AI technologies in many states, and the Australian framework gives explicit permission to use generative AI. Every sector has the opportunity for equal participation in understanding how best to develop these technologies.

Limiting students’ experience of generative AI to just chatbots in traditional classroom contexts, trained on equally traditional curriculum material, is not how we’re going to achieve this. We should take a leaf out of vocational education and explore ways that these technologies can be used to get the job done.

If you’d like to get in touch to discuss GenAI, PD, or consulting, please use the form below: