Use of Artificial Intelligence in Sixth Form Education: Position Statement

By Craig McDonagh, assistant principal - St Dominic's Sixth Form College and Matt Jones, digital lead - Blackpool Sixth Form College

This position statement is intended as a guiding document for stakeholders involved in the integration of AI in sixth form college education in the UK. It calls for a balanced approach that leverages the benefits of AI while addressing potential challenges and ethical considerations.

Introduction

Until very recently, generative artificial intelligence was science fiction. On November 30, 2022, the technology company OpenAI released a large language model (LLM) which brought artificial intelligence into the real world and to the general public. ChatGPT-3's capabilities were unlike anything else, helping it to reach one million users in five days. Since then, OpenAI have released two more advanced GPTs (with more on the way) which currently have over 180 million users.

Google, Meta, and X (formerly Twitter) have since released their own GPT competitors, along with a host of other start-up technology companies. New AI companies are forming at an accelerated rate, with almost 1300 new AI companies created in the UK in the 12 months following November 30, 2022.

This explosion of AI companies has resulted in a similar increase in use cases. For students, AI can be a wonderful personalised teacher, patiently explaining any question asked of it. It can assess, identify areas for development, and re-test, and of course, it can do homework. For teachers, AI can now plan your lessons, create your resources, and even mark that pile of homework it completed earlier.

AI is already having an impact in our classrooms and colleges, and with more sophisticated and capable models imminent, this impact is likely to increase in the coming years. As a sector, how we face these changes should reflect our commitment to innovation and new opportunities for learning. It should reflect a dedication to preparing our students for a future where technology plays a vital role.

Current Position

AI adoption is varied across the sector, for many reasons: expertise, confidence, interest, and the availability of human resources. Some colleges have created AI policies and have adopted strong working practices while other colleges are yet to start utilising AI at all. External guidance from various bodies is yet to offer a clear vision or future direction, leaving colleges to define how AI can and should be used for themselves. The sector would benefit from a more systematic approach to AI development and utilisation.

Continued Professional Development

The growth in the use of AI in education has been mostly organic to date, fuelled through the personal interest of a small number of enthusiasts. As such, the development of best practices and exploration of use cases has occurred mostly independently; this inevitably leads to replication of work and an inefficient use of resources across the sector.

This suggests a centralised approach to AI and training should be developed for the sector. Centralised professional development would allow for feedback loops to determine with greater certainty which of the plethora of tools available add the most value, specific to the needs of our learners.

What makes AI and LLMs different from other recent innovations, such as VR and AR, is the rate of adoption, facilitated by their low barrier to entry. Anyone with a smartphone and an internet connection has access to LLMs, such as ChatGPT, which can do things that were impossible only a short time ago.

Generally, young people adopt new technologies more quickly; it is likely that most classrooms already have some students utilising AI. Without appropriate training across the sector, untrained teachers may soon find themselves in the undesirable position of being the only person in the classroom unfamiliar with these new tools.

Therefore, it is important that teachers have at least a basic understanding of how an LLM functions. Being unaware of the relative strengths and drawbacks of such tools will be a barrier to the education of students who will look to teachers for guidance. Without a level of awareness within the teaching body, this guidance will fall short.

Educating Students on AI

In addition to training for teachers, students must be educated on the safe and appropriate use of AI. A sector-wide framework for the use of AI should be implemented to pre-emptively counter the potential risks of widespread AI adoption.

Currently, the most pressing risk for many in the sector is likely to be plagiarism. One can easily see why some students are tempted to use AI to complete homework or coursework. As teachers, we need to find a pragmatic balance between allowing students to utilise AI for their learning so that they can complete a task, without accepting work completed by AI tools. As part of this AI education, teachers may need to go back to basics, explaining how homework and coursework fit into their learning journey and contribute to preparation for public exams.

AI has enormous potential to support learners: ideation, personalised feedback, SPAG reviews, reframing content, and generally being an endlessly patient tutor are just some of the ways that students are already enhancing their learning with AI. Making students aware of these capabilities and supporting them in their use should be part of our approach to AI.

AI models are far from infallible, as demonstrated through their "hallucinations" — the term used to mean the incorrect or entirely fabricated responses sometimes provided by models when asked a question. This happens because the models do not intrinsically understand content; rather, they predict what they believe the next word or sentence should be based on patterns observed in training data. This, at its source, is the same reason that predictive text sometimes provides bizarre suggestions. Ensuring that students understand the limitations of AI models will be a vital part of AI education.

Ethical Considerations

There is no consensus view of the potential impact of AI advancements on society, even among the experts working in the space. As such, it is unrealistic to expect a homogeneity of opinion on its use in our sector.

The ethical use of AI, especially the idea of creating general AI (AGI), raises serious concerns. Developers must be sure they can control AGI such that it follows our values. Many in the field have predicted the creation of AGI by the end of the decade. If AGI becomes smarter than humans, some experts worry that it could be impossible to control or even act in harmful ways. Although beyond their control, educating students about the potential futures of AI should be part of the conversation.

The technology gap is well documented within our sector. For AI, the barrier to entry is low but rises for the most powerful tools in the space through companies' use of premium subscriptions. This is something that colleges may need to consider with widespread use of AI models on the horizon.

Colleges need to keep pedagogy at the heart of decision-making, being careful not to become swept away with new ideas. One illustrative example is the promise of highly personalised learning tools. While these tools might increase progress in some aspects of education, moving away from collaborative learning and soft-skill development has significant downsides for holistic student development.

Research in a Rapidly Changing Landscape

The rapid advancements in AI, especially LLMs, present a challenge in keeping educational policies and practices up-to-date. This situation calls for continuous research and adaptability in the education sector. Traditional curriculums and assessment methods are not keeping pace with emerging technologies, despite a strong interest in using AI to solve problems faced by staff and learners. However, the immediate demands of teaching often leave little time for the necessary deep thinking, research, and reflection needed for the safe and effective integration of AI.

To address this, resources must be allocated for research and experimentation in AI. This includes upskilling staff to adopt new technologies. The potential of AI, such as training context-specific bots with ethical responses, remains largely untapped due to our limited capacity for experimentation. This could create a cycle where avoiding risks leads to stagnation and missed opportunities to establish best practices.

We should develop mechanisms for ongoing research into the latest AI developments and for sharing best practices among leaders and teachers. There should be an open dialogue with government bodies to ensure that practical insights inform national planning and processes. Relying on individual efforts to integrate AI tools at varying rates is inefficient and unlikely to yield the best outcomes. Therefore, it is crucial to collect, review, and systematically apply collective learning to enable effective and informed progress across the sector.

By adopting a structured approach to AI integration, the education sector can ensure that all institutions benefit from advancements, leading to better teaching and learning outcomes.

Conclusion

Given the rapid and ongoing advancements in AI technology, it is challenging to define a single best approach to AI in education. The broader debate around AI is unlikely to reach a clear conclusion soon, and regulatory guidance will continue to evolve. Therefore, it is essential for colleges to adopt a flexible and proactive approach:

Upskill Staff and Develop AI Literacy: Ensure all staff have a basic understanding of AI, including ethical considerations, academic integrity, and safety concerns. Training should cover what a Large Language Model (LLM) is, effective techniques for using AI (promptcraft), and the importance of critically evaluating AI-generated outputs.

Stay Informed on AI Developments: Keep up with the latest advancements in AI and the surrounding research to remain current and informed.

Reflect on AI Use by Students: Consider and assess the impact of AI on student learning and make thoughtful decisions regarding its integration.

Collaborate and Share Best Practices: Work together across the sector to share findings, strategies, and emerging good practices to enhance collective knowledge and effectiveness.

Adhere to JCQ and OFSTED Requirements: Stay aware of existing and evolving guidelines from regulatory bodies to ensure compliance and best practice.

Communicate Transparently with Stakeholders: Provide clear guidance to parents and students on positive AI use, highlighting both the benefits and potential risks of misuse.

By taking these steps, colleges can position themselves to respond effectively and with agility to the opportunities and challenges presented by AI, ensuring that they are prepared for the future.