Cryptic compiler error messages continue to present a significant barrier for novice programmers, especially in foundational languages like C. Although large language models (LLMs) can generate accurate and comprehensible error explanations, their computational requirements, propensity for over-assistance, and privacy concerns constrain their suitability for widespread adoption in educational tools. This work investigates how Supervised Fine-Tuning (SFT) can enhance the performance of smaller, open-source models when explaining C compiler errors to students in introductory programming courses (CS1/2). We derive a training dataset of 40,000 input-output pairs from CS1/2 student C compiler errors to fine-tune three open-source models: Qwen3-4B, Llama-3.1-8B, and Qwen3-32B. Model performance was assessed through a dual evaluation framework involving expert human reviewers and a large-scale automated analysis of 8,000 responses using an ensemble of models as judges. Our results indicate that SFT significantly improves both expert and LLM-as-judge ratings in smaller open-source models, with reduced gains in the larger model. We analyse the trade-offs between model size and quality, and validate LLM-as-judge by demonstrating inter-rater agreement with experts. Our findings demonstrate that fine-tuning smaller models on high-quality data is a viable strategy for creating specialised pedagogical tools. We provide a replicable methodology for enabling broader access to advanced AI capabilities within educational contexts, especially with smaller, economical models.

Lengthy lectures can overwhelm a student's cognitive capacity, particularly in CS1 courses, where students often struggle with foundational programming principles. Bytesized is a lecture snippet tool that curates short, targeted videos from lecture recordings in response to student queries. We deployed Bytesized in a CS1 course in C, which revealed a high engagement with the tool as a way to supplement the consolidation of core concepts. The early results suggest that tools such as Bytesized are key in providing CS1 students with the necessary support to develop effective study strategies to consolidate fundamental programming teachings.

As artificial intelligence (AI) reshapes industries and society, there is an increasing demand for early AI education to ensure future generations are equipped with the skills and critical thinking required to navigate an AI-driven world. Day of AI Australia is part of an international initiative designed to equip educators and students in upper primary and early secondary school (ages 10 to 16) with foundational AI literacy. The Australian program provides a series of lessons that introduces students to core AI concepts such as machine learning, natural language processing, and AI ethics. Insights from post-lesson feedback data from over 60 participating school teachers highlights the efficacy of the curriculum, the challenges faced, and the potential for scaling AI education in Australian classrooms. We provide practical recommendations for integrating AI literacy into national and global curricula, with the goal of preparing students for the AI-driven future.

As artificial intelligence (AI) continues to transform industries and impact daily life, it's crucial for K-12 students to be AI literate. However, many teachers feel unprepared to teach AI-related concepts. This session will focus on strategies to equip teachers with the knowledge and confidence they need to bring AI into their classrooms. Drawing on insights from the Day of AI, which has provided AI education to half a million students and professional development to thousands of teachers worldwide, the session will explore successful models for scaling teacher training in AI. Topics will include debunking common misconceptions about AI, addressing logistical challenges, and offering practical, adaptable resources that allow teachers to integrate AI into a variety of subjects. Attendees will have the opportunity to share their experiences and insights, particularly focusing on how to overcome barriers to teaching AI, such as lack of prior technical expertise or insufficient time to devote to professional development. The session will emphasise the importance of creating a collaborative and supportive network for developing and sharing AI teaching resources. Through interactive discussions, participants will collaboratively explore strategies to expand AI literacy and develop teacher confidence, aiming to create a sustainable model for equipping educators to teach AI at scale. This session will foster community engagement, promote shared learning, and serve as a launchpad for continued collaboration on AI education.

Formative feedback spans various domains, from education to businesses and creative endeavours. In educational contexts, feedback enriches students' learning and work quality through reflection. However, providing effective feedback at scale is challenging. Students struggle to engage with feedback, often due to lack of feedback literacy. Recent advancements in Natural Language Processing, a branch of Artificial Intelligence, provides opportunities to evaluate how we can support feedback providers in its quality and scale. This poster paper presents an overview of key feedback challenges, attributes of high quality feedback, and introduces the Intelligent Feedback Assistant (IFA), an innovative NLP-based system designed to assist educators in delivering high-quality feedback. IFA operates as an ensemble of machine learning models and non-AI systems to guide educators in refining their feedback, ensuring it embodies attributes of effective feedback - actionable, specific, justified, and positive. IFA is supportive, not generative, ensuring the feedback provider remains central to the feedback provision process. The tool design, and outcomes of IFA offers a promising path for scaleable, high-quality formative feedback in education and beyond.

Students in introductory computing courses often lack the experience required to effectively identify and resolve errors in their code. For such students, Programming Error Messages (PEMs) are often the first indication of an error, and could provide valuable debugging guidance. However, in many cases, such as with standard C compiler implementations, PEMs are largely unsuitable for novices. Confusing, misleading, and filled with terse language and jargon, these messages instead act as an additional source of difficulty.

In the challenging field of introductory programming, high enrolments and failure rates drive us to explore tools and systems to enhance student outcomes, especially automated tools that scale to large cohorts. This paper presents and evaluates the dcc --help tool, an integration of a Large Language Model (LLM) into the Debugging C Compiler (DCC) to generate unique, novice-focused explanations tailored to each error. dcc --help prompts an LLM with contextual information of compile- and run-time error occurrences, including the source code, error location and standard compiler error message. The LLM is instructed to generate novice-focused, actionable error explanations and guidance, designed to help students understand and resolve problems without providing solutions. dcc --help was deployed to our CS1 and CS2 courses, with 2,565 students using the tool over 64,000 times in ten weeks. We analysed a subset of these error/explanation pairs to evaluate their properties, including conceptual correctness, relevancy, and overall quality. We found that the LLM-generated explanations were conceptually accurate in 90% of compile-time and 75% of run-time cases, but often disregarded the instruction not to provide solutions in code. Our findings, observations and reflections following deployment indicate that dcc --help provides novel opportunities for scaffolding students' introduction to programming.

In supporting online student cohorts, we experienced challenges in achieving the same quality of engagement asynchronously as we do through face-to-face discussions in a classroom setting. The educational model we use is based upon students progressing through weekly tasks designed to support development, and measure achievement of learning outcomes. In this model, once a student has completed a task, interactions between the student and instructor provide feedback to the instructor of the student's understanding. The feedback system gives confidence that the student understands their work and aids in identifying learning intervention opportunities. To achieve this asynchronously, we developed and integrated an audio-discussion tool known as Real Talk into our Learning Management System (LMS). The tool allows instructors to record discussion prompts tailored to a student's completed task and has the LMS replay the prompt(s) and immediately capturing the student's response. These interactions replicate essential aspects of face-to-face, in-person discussions by not affording the student opportunities to research and rehearse responses, which we previously experienced when using asynchronous discussion or quiz tools for this purpose. In this paper, we present the implementation of the Real Talk tool and discuss results evaluating how effective it was at allowing instructors to identify opportunities for learning interventions in introductory computing courses. The results confirmed that the tool has assisted in identifying knowledge gaps not identifiable in students' submissions alone.