Generative AI in further education and skills – the myths, the threats and the opportunities

Michael Webb, director of technology and analytics at Jisc’s national centre for AI

It may seem like generative AI has been a hot topic in education for quite a while now, but it was only with the introduction of OpenAI’s ChatGPT at the end of 2022 that the conversation really started to ramp up.

Generative AI refers to a type of artificial intelligence that can create new content such as text, images or media based on existing data. Most forms of generative AI, including ChatGPT, are trained using vast amounts of data from the internet and other – unspecified – sources.

From an educational perspective, the rapid emergence and the wider implications of this technology may appear daunting. The key thing is not to panic, but to embrace the opportunities while understanding the challenges.

So, what is the best approach for colleges to take when it comes to generative AI?

Be aware of the myths

When something evolves at such a rate as generative AI, there are bound to be areas of the unknown or the misunderstood.

Firstly, banning its use is not an option: it’s already integrated into our daily work processes through Google, Microsoft Bing and other widely used apps.

And, despite all the media hype, ChatGPT is not the only generative AI tool out there. There are multiple examples of tools using this type of technology that can develop content including writing, coding and imagery.

It’s also important to be aware that generative AI tools don’t know it all. For example, some – such as the basic version of ChatGPT – have no access to external data such as the internet and so can only draw upon the material on which they were trained. And, in the case of ChatGPT, this has a cut-off date of September 2021.

Another common misperception is that AI can “understand” the question being asked and the answer it produces. The reality is that these tools simply use their pre-training to predict the most likely next word in a given sequence. This can result in inaccuracies and highly plausible untruths, so any output requires rigorous checking.

When it comes to plagiarism, relying on AI detection tools won’t solve the problem either. While these are being developed almost as quickly as the technology is evolving, they are not foolproof and no system today can conclusively prove text has been written by AI.

Those placing all their eggs in the AI detection tool basket run the risk of missing instances of AI-generated content or of falsely accusing students of cheating.

Understand the threats

Without greater understanding of how generative AI works and what it can (and can’t) do, there is a risk that educators and students alike may be using tools that are not fit for purpose, or not using them to their best advantage.

And, as the creators of these tools work out how to monetise them, the threat of increased digital inequality looms.

ChatGPT is free (for now) but its successor ChatGPT Plus, based on the more advanced GPT-4 LLM (large language model), is not. Nor, obviously, are any of the more focused tools which are coming onto the market.

This could widen the digital divide if some learners are unable to afford to access the same AI capabilities as their peers.

Embrace the opportunities

AI undeniably has great potential to alleviate one of the greatest challenges currently faced by teachers in FE: the ever-increasing workload.

For example, AI-based tools like TeacherMatic that are specifically designed for use in education really can make teachers’ lives easier.

By completing routine tasks such as lesson planning, schemes of work and resource creation in a fraction of the time it would otherwise take, they allow teachers to prioritise education over admin and help them to be more effective in the classroom.

Equipping staff with a clear basic understanding of how these tools work and giving them the time and training to experiment for themselves can help lessen the load and boost creativity.

The way forward

As these tools proliferate and regulation for ethical development and usage is enacted, the opportunities to harness generative AI to improve teaching and learning will only increase.

It’s clear that the key to successful use of AI tools in further education is improving the knowledge of generative AI for both staff and learners.

Not all colleges are at the same level of understanding when it comes to generative AI and it’s the responsibility of organisations such as Jisc to help bridge that knowledge gap to upskill staff and limit the risk of students being left behind.