Using AI in the Classroom - Friend, not Foe

By Anam Javed
Victorian Academy of Teaching and Leadership

Artificial intelligence (AI) dominated the news in 2023, as tools like ChatGPT seemed to propel AI from science fiction into everyday life. In response, every aspect of society is now trying to determine what the advent of AI will mean for them. Here, we take a look at how educators can harness AI to meaningfully improve the education experience.

What is generative artificial intelligence?

AI is a branch of computer science that aims to create systems capable of performing tasks that include learning, reasoning, problem-solving, and data analysis.

Generative artificial intelligence (GAI) is a subset of AI, which specifically focuses on generating content, be it in the form of text, images, music, or other media. While traditional AI is often about analysis and decision-making based on existing data (like recognizing patterns or making predictions), generative AI is about harnessing creativity and creating dynamically evolving content.

Chatbot Generative Pre-trained Transformer, or ChatGPT, is an AI model developed by OpenAI that took the world by storm in late 2022.¹ It is part of the GPT series, which are large language models (LLMs) designed to understand and generate human-like text based on the input they receive - a prominent framework for GAI.

It is important to note the limitations of GAI models like ChatGPT. They can only assemble responses based on the data they have been trained with, meaning the quality and depth of responses is only as good as the training data. They also aren’t able to access real-time information, but may be periodically ‘updated’ with new data - comparable to using a textbook published every few years.

What is prompt engineering, and why does it matter?

Even if you haven’t tried ChatGPT, you have likely engaged in some form of AI. Whether it be the pop-up help tool on a website, or the automated service that asks you to describe the reason for your call, you’ll know how important it is to phrase your question so that you are given the answer you need.

Prompt engineering is the skill of crafting questions or instructions (prompts) to effectively communicate with an AI system, ensuring that the AI understands and responds in the most useful way. It is a critical skill for interacting with AI models, particularly those utilised in chatbots or content generation tools.

Well-engineered prompts harness the full capabilities of an AI, leading to more accurate, relevant, and valuable responses. Clear, specific prompts minimise the chances of misinterpretation by the AI, reducing irrelevant or incorrect responses. Effective prompts improve the interaction between the user and AI, making it more efficient and user-friendly.

What does a good prompt look like?

While simple prompts can yield decent results, the quality of the output hinges on the amount and nature of information you include in your prompt.

Instruction: The primary question or task you’re asking the model to perform.
Context: Additional background information that helps the model understand the query better.
Role: A role assigned to the LLM.
Inputs: Data or parameters that the model might need to generate the desired output.
Output: Examples of what you want the model to generate, and in what format.

Now, let us synthesise this into an example prompt:

You are a year 10 student in charge of creating a four-episode podcast series to be released over four weeks at your school, focusing on the use of AI to write speculative fiction. Create a list of steps to follow to create this podcast series, with a progressive outline for each episode, including a topic for each week. Each episode is to be 30 mins long and is relaxed and conversational in nature. Generate the episode outlines as a table.

Applications of GAI in a science classroom

AI is becoming a pivotal tool in transforming learning experiences, especially in science classrooms. AI's potential to personalise learning, provide real-time feedback, and create immersive educational experiences is unparalleled.

The table below outlines how AI can be used in a science classroom to make learning more accessible and engaging:

GAI can also be used to promote interdisciplinary learning with science as a central focus. Consider the following prompt, which can be fed into the freely-accessible ChatGPT 3.5:

Use NASA’s Exoplanet resource to create a fictional planet beyond our solar system where humanoid creatures thrive. Consider elements like the terrain, weather, and gravity, and how they impact the type of species that can survive on this planet. Use Future Timeline to describe sociopolitical issues impacting this planet if it was to exist 500 years from now.

Note that this particular prompt works because NASA’s Exoplanet resource³ and the concept of a Future Timeline⁴ form part of ChatGPT’s training data.

Next, you can use Craiyon⁵ to generate some images of this fictional exoplanet, using the following prompt:

Generate an image of a fictional exoplanet densely populated with pink foliage, and surrounded by a dense layer of purple gases, in the style of Van Gogh’s ‘The Starry Night’.

Victorian teachers harnessing GAI in the classroom

Teachers' views on AI in education are diverse and multifaceted, reflecting a blend of optimism, caution, and curiosity.

Many educators see AI as a tool to enhance the teaching and learning processes, not as a replacement for teachers but to complement their skills. With the release of the Australian Framework for GAI in Schools,⁶ school teachers and leaders are now being gradually equipped with knowledge and frameworks that guide the responsible and ethical use of generative AI tools in ways that benefit students, schools, and society.

There is apprehension about the ethical use of data about and created by students, with teachers emphasising the need for robust data privacy and security measures. The eSafety Commissioner has published a range of resources to mitigate risk associated with the use of AI in schools.⁷

The following are some ways in which educators can bring students onboard with using AI ethically, collaboratively, and effectively to boost learning and engagement in the classroom, across all disciplines:

‍Awareness and Education:

Create a shared, agreed upon school-wide framework for the use of AI tools. Raise community awareness by running workshops that enlighten and empower teachers, students, and parents about the ethical use of AI tools like ChatGPT, along with academic integrity and its implications. Through this lens, educators can clarify what constitutes cheating and its consequences. This can be used as a culture-building opportunity, and hence encourage a culture of honesty and integrity, linked to school and community values.

Adapt Assessment Design:

Modify the design of assessments to reduce the likelihood of cheating. As much as possible, weave in opportunities to co-design assessment tasks with students. This could include:

• Implementing project-based assessments where students must demonstrate their learning process over time.
• Incorporating oral exams or presentations where students need to explain their understanding verbally.
• Considering alternative forms of assessment like portfolios, lab work, or practical demonstrations.

Encourage Creativity and Personalisation:

Assign formative assessment tasks that require the use of student voice and creative thinking, as these are harder to replicate using AI.

Peer Review and Collaborative Work:

Implement peer review systems where students evaluate each other's work, promoting accountability. Students are often very well-equipped to recognise and honour each other’s work.

Teacher Involvement and Follow-ups:

Have teachers engage in an ongoing manner with students' work, asking follow-up questions or having discussions about their submissions, with this being reflected in assessment rubrics.

The key is to create an educational environment where learning is valued over grades, and where students understand the importance and benefits of honest academic practices.

Where to next?

With a range of generative artificial intelligence platforms on the market, ongoing professional learning for teachers will have an increasingly significant role to play in making sure that no teacher gets left behind.

Organisations such as the Australian Centre for the Moving Image (ACMI) and Digital Learning and Teaching Victoria (DLTV) are playing a pivotal role in connecting educators across the state via the sharing of resources and professional learning, while peak curriculum bodies such as Victorian Curriculum and Assessment Authority (VCAA) are providing targeted support to teachers to leverage resources with regards to curriculum design and implementation.^8,9,10 The Department of Education Tech Schools are currently offering a range of interdisciplinary programs to partner schools which harness the power of artificial intelligence to embed interdisciplinary knowledge and skills into learning and assessment.¹¹

The Teaching Excellence Program¹² offered by the Victorian Academy of Teaching and Leadership (VATL)¹³ is one such avenue, featuring a year-long professional learning program curated to connect teachers to like-minded peers, while upskilling them using evidence-based approaches, including in the field of generative artificial intelligence. Teachers are equipped with the skills and resources to undertake authentic inquiries where they explore problems of practice and collaborate with peers to come up with viable and transformative solutions and learnings. Once teachers graduate, they are able to join the Alumni Network, which offers alumni the opportunity to engage in professional learning and networking events. Alumni also have access to an online community where they can engage in discussions, share resources, and support one another.

The integration of AI in science education offers an exciting frontier for enhancing both learning and engagement. By personalising education, providing interactive learning environments, and revolutionising assessment methods, AI not only makes science education more effective but also more engaging for students in the long run. As we continue to explore the capabilities of AI, the potential for further innovations in the educational sector remains boundless.

—

‍Anam Javed is the Master Teacher for Technologies at the Victorian Academy of Teaching and Leadership. Anam is passionate about harnessing the potential of digital technologies to combat social inequity and improve educational outcomes for students from all backgrounds.

References:

1. ChatGPT. chat.openai.com
   
2. History of Artificial Intelligence. (2019, January 30). UQ. qbi.uq.edu.au/brain/intelligent-machines/history-artificial-intelligence
   
3. NASA. (2015, December 17). Exoplanet Exploration: Planets Beyond our Solar System. NASA. exoplanets.nasa.gov
   
4. One example of a future timeline is made by William Fox. futuretimeline.net/21stcentury/2050-2059.htm
   
5. Craiyon, Formerly DALL-E Mini. craiyon.com
   
6. Clarke, M. (2023). Australian Framework for Generative Artificial Intelligence (AI) in Schools. Department of Education. education.gov.au/schooling/resources/australian-framework-generative-artificial-intelligence-ai-schools
   
7. Generative AI – position statement | eSafety Commissioner. (2022). ESafety Commissioner. esafety.gov.au/industry/tech-trends-and-challenges/generative-ai
   
8. ACMI - Australian Centre for the Moving Image. acmi.net.au
   
9. DLTV - Digital Learning and Teaching Victoria. dltv.vic.edu.au
   
10. Digital Technologies - Victorian Curriculum. VCAA. victoriancurriculum.vcaa.vic.edu.au/technologies/digital-technologies/curriculum/f-10
    
11. Victorian Tech Schools. vic.gov.au/tech-schools
    
12. Teaching Excellence Program. VATL. academy.vic.gov.au/initiatives/teaching-excellence-program
    
13. Victorian Academy of Teaching and Leadership (VATL). academy.vic.gov.au

‍

‍

‍