Web Content Display (Global)

Guidelines for using Artificial Intelligence (AI) in SACE assessments

Students have always been keen to use new and exciting tools to support their learning. Public access to generative AI through ChatGPT has given students access to a powerful tool to aid their studies, including when completing school-based submitted assessments and external investigations. This access will only improve as AI tools are integrated into the websites and software that students use every day. AI technologies have already had a profound impact on education and student outcomes, improving student access to learning and assessment through tools like text-to-speech, and in supporting schools to manage and support student learning through tools such as Learning Analytics.

This page outlines the SACE Board’s approach to generative AI in SACE school-based assessments and submitted investigations. Central to this approach is consideration that the work submitted for assessment is the student’s own, and that any sources used are acknowledged. Guidance provided aligns to the existing suite of assessment and academic integrity policies of the SACE, relying on a shared partnership between schools and the SACE Board to ensure the integrity of school-based and external assessments.

These guidelines should be read alongside the Supervision and Verification of Students’ Work Policy and Procedure [PDF 180KB], which is underpinned by the principle that students must only submit work "that is their own, produced without undue assistance from other people or sources." The breach of rules page provides more advice for schools about the rules for students who undertake SACE assessments and the procedures to follow when a breach of rules occurs. 

Can students use generative AI in their school-based assessments and investigations?

Yes.

Generative AI tools, including ChatGPT, are like the many other information sources available online. They are a research tool that can provide access to knowledge for students.

The SACE Board does not specify which ways of accessing knowledge are valid across subjects. We rely on the discipline expertise of educators to teach students how to evaluate the many sources available to them, and which are appropriate for use in their disciplines. We encourage broad research from a number of sources and for students to always view sources critically, as we know their teachers have taught them to.

Back to top


How much AI can be used in a submitted task?

There is no limitation on the amount of information a student can gather from a generative AI source to use in an assessment task. Teachers should work with students to ensure that they understand the limitations of AI sources, and how the overuse of limited sources can impact the quality of their work.

Back to top


What ways can AI be used in a submitted task?

There are many appropriate ways that students can use generative AI in their submitted assessment tasks. Students can access generative AI sources to research and inform their assessment, just as they would for a textbook or other traditional information source. As with these traditional sources, students cannot use text from generative AI and present it as their own.

Students may also use generative AI tools to support their assessment development and writing. These tools are increasingly integrated into software used by students, and at times students may be using them without them realising that they are (including predictive writing support provided by many word processors). While the use of these tools is appropriate, students must ensure that the work submitted for assessment is their own. These tools can only be used to support a student’s own writing processes, not to replace them.

For example, it may be appropriate for a student to use an app to provide suggestions that they can consider to improve their writing during their drafting process. However, it is not appropriate for students to enter their draft into an app which could change the syntax and structure of the text without the students making decisions about phrasing.

Students must not submit work generated by AI as their own work.

Back to top


How can students acknowledge their use of generative AI?

In all contexts, students must provide an acknowledgement of any generative AI used as a part of their task. It is expected that students will acknowledge any use of AI in a way that is appropriate for the subject and school context. This acknowledgement should declare which tools were used and provide a list of all prompts that were entered to generate any information for the task. This practice is particularly useful for tasks where individual sources are not directly referenced throughout, or where the AI provided broader support of the student work. In some cases, such as image generating AI, providing the output images generated and any reference images entered into the tool, would also be appropriate.

In some cases, it is also appropriate for students to make specific references to AI generated work when used throughout their task, as they would when citing other information sources. In most cases, this would include students providing a reference to work created by generative AI when quoted or paraphrased in their task including: the name of the AI tool used, a link to access this resource (if appropriate) and any prompts that were entered to generate the response. 

The SACE Board Guidelines for Referencing [DOC 160KB] have been updated to include some suggestions of how schools and students might choose to reference generative AI.

Back to top


How can I use AI in my assessment task design?

Educators across schools and tertiary providers have begun to consider how they might utilise AI in task design or their assessments. Secondary schools across South Australia have already begun to consider how they might assess students in light of access to AI tools, and how this technology can be leveraged to support their students to grapple with and understand content. Considering how learning design and pedagogy can be supported with access to AI – as well as questioning the what and the how of student assessment – is emerging as a positive way to navigate the uncertainties that this (and other) new technologies pose. Some emerging suggestions include:

  • integrating working with AI tools into task design, requiring students to generate initial ideas or improve AI outputs as a starting point for tasks
  • analysing prompts used in the assessment task with AI tools and evaluating the response the tools can generate
  • focusing assessment on the processes and skills used to create an outcome
  • developing assessment tasks that AI tools do not, or cannot, know about including self-reflection, responding to classroom discussion prompts and embedding student experience into tasks.

Back to top


Can students use generative AI tools in their external exams?

No.

The use of ChatGPT and other generative AI tools is not appropriate for use in external exams. For 2023, the security settings for the SACE Exam Browser have been raised to the highest setting, and additional testing will continue to verify that generative AI tools are blocked in e-exam settings. 

Verification and trust of assessment is a shared partnership between students, schools and the SACE Board. Exams are an extension of the trust and verification in established policies and practices. Raising the security settings of the e-exam platform is one method that will support existing academic integrity policies and the active invigilation of examinations.

The SACE Board is re-engaging and collaborating with IT experts in schools to ensure clear communication and best practice across the system. We are also reviewing our e-exam support materials to best prepare schools, invigilators and students for exam day, and how best to respond to anything unexpected or out of their control. 

Back to top