The University is committed to providing you with the tools, skills and capabilities you need to thrive in your studies and the future of work. ChatGPT Edu is intended to meet the needs of our students, whether learning, studying or researching. Every day, we see our students exploring and experimenting with AI more and more. We need to respond to that growing curiosity and provide you with the appropriate tools, training and support so that you can use AI in secure and responsible ways.
It means fair and equitable access to AI across our institution and as aligns with the University's principles and guardrails for the responsible use of AI. ChatGPT Edu operates in a University-managed environment, with built-in safeguards designed to protect our data.
We selected ChatGPT Edu because it offers the strongest overall combination of capability, usability, institutional scale, governance and value for money. Network logs from across the University also showed that their free tool was by far the most widely used AI tool, demonstrating strong existing adoption and familiarity among staff and students.
The platform provides an intuitive and mature generative AI experience that can support learning, teaching, research and University operations. Its uses include drafting and summarising content, analysing information, supporting coding, conducting advanced research and exploring ideas through natural-language conversations. Importantly, it can be made available across the University through a single, centrally managed environment rather than relying on fragmented consumer tools and individual subscriptions.
In addition to the existing tools available in our AI Inventory, ChatGPT Edu has undergone our thorough procurement and evaluation process that considered multiple vendors and solutions. The tool has been reviewed and endorsed in line with our Responsible AI Framework.
The University recognises that AI is becoming an increasingly important part of education, research and the workplace.
You will have opportunities to build the skills and confidence needed to use AI productively, responsibly and ethically, while continuing to apply critical thinking, academic integrity and human judgement in your study. AI is intended to assist people, not replace their expertise, decision-making, compliant practices or learning.
The University of Sydney's partnership with OpenAI to provide AI access to all students and staff is one of the largest AI investments in Australian higher education. While the cost is commercial in confidence, the investment reflects the University's broader commitment to our people – giving all students and staff equitable access to advanced AI tools, alongside the training, safeguards and support needed to use them effectively, responsibly and with confidence.
A student pilot of ChatGPT Edu is planned for October 2026, around the time of the staff release of the tool. Following the pilot, ChatGPT Edu is scheduled for release to all students in Semester 1 2027.
We are taking a phased approach to rollout, enabling staff first so we can gather insights, refine support resources and ensure a smooth experience before extending access to students. Further details, including timing, access arrangements and support resources, will be communicated closer to launch. You will receive updates through the University's usual communication channels as plans are finalised.
ChatGPT Edu will offer features and functionality beyond what is available in the free version of the tool.
At launch, users will have access to the University-approved baseline version of ChatGPT Edu, including the latest GPT models, conversational chat, shared project spaces for collaboration and image generation capability.
Additional capabilities such as custom GPTs, agents, plugins, connectors and other advanced features will be individually assessed before being enabled. A high-level feature roadmap will be published shortly.
No. ChatGPT Edu will be one of the University's endorsed AI offerings. The University will continue to assess and invest in other AI tools where they are appropriate for specific use cases, disciplines, research activities or business needs.
The tool complements our homegrown and award-winning tool Cogniti, which will remain the primary tool for educators and students in teaching and learning activities, as well as other tools in use at the University, including Harvey AI at the Sydney Law School.
The University provides access to a range of AI tools, each designed for different purposes and user groups. As a general guide, Cogniti will remain the University's primary AI platform for teaching and learning activities in the classroom, while ChatGPT Edu can supplement your study and student activities more broadly, including outside the classroom. Microsoft's Copilot Chat will remain as part of your Office application suite.
If you're unsure, ask your educator, as they will be able to provide guidance and help you select the most appropriate tool for your needs.
ChatGPT Edu, accessed through the University's single sign-on (SSO), should be used for University work and study where it has been provided. Data entered into ChatGPT Edu does not train publicly available large language models (LLMs) by default, and access is managed through University authentication.
Personal ChatGPT accounts, including free and paid versions purchased individually, are not managed by the University. Only information that is already publicly available should be entered into these services. Protected, highly protected or internal University information must not be entered into public AI tools.
For detailed guidance on the responsible use of the tool, we strongly recommend you review and stay up to date with our ChatGPT Edu specific AI guardrails. These will be made available in line with the student launch of ChatGPT Edu.
While the University of Sydney's version of ChatGPT does protect your data, some data and use cases are not permitted.
The University of Sydney's ChatGPT Edu workspace has Australian data residency at rest and globally routed inference. This means that while conversations, uploaded files and other information retained by the service are stored within Australia, prompts and uploaded content may be processed transiently outside of Australia to generate responses. The service remains subject to the University's governance, privacy and security requirements.
Your chats, including uploaded and generated content, are not visible to other staff or students. However, authorised University administrators may access or export records for security, legal, governance, retention or investigations.
Information submitted to ChatGPT Edu won't be used to train any of the vendor's models, and the enterprise environment includes identity and access management, encryption, audit logging and configurable data-retention settings. This means that user data will remain secure and private, provided users are following the University's specific AI guardrails for ChatGPT Edu.
The University closely considered users' privacy and data security through the privacy, cybersecurity, architecture and responsible AI reviews that we completed as part of the procurement process, with higher-risk features like external connectors and autonomous agents disabled until separately approved.
The rollout is governed by our Responsible AI Framework and phased with formal checks. Protecting University data is a shared responsibility. While these technical and governance controls help reduce risk, they cannot replace good judgement. Users will remain responsible for following the University's specific AI guardrails for ChatGPT Edu and for ensuring that they comply with University policies and applicable legal requirements.
We will provide the specific AI guardrails for ChatGPT Edu ahead of the tool launch and incorporate these into all system training, support materials and communications.
As a research-intensive university that is on the forefront of discovery, we have a responsibility to ensure that the tools we provide our researchers will not inadvertently compromise their competitive advantage in fast-moving disciplines, provided they are using the tools in accordance with University guidance. The conversations we have had with OpenAI and the due diligence we have undertaken through our privacy, cybersecurity, architecture and responsible AI reviews give us a degree of confidence that user data will remain secure and private.
Security, privacy and responsible AI were central to our review of ChatGPT Edu. Information submitted to the platform will not be used to train their foundation models. The enterprise environment includes identity and access management, encryption, audit logging and configurable data-retention settings.
Privacy, cyber security, architecture and responsible AI reviews supported proceeding with the agreement, subject to the agreed controls being completed before broader release. Higher-risk capabilities, including external connectors and autonomous or write-capable agents, will remain disabled until separately assessed and approved. The solution also supports a controlled approach to adoption.
Responsible AI is a key consideration in the rollout of the University's enterprise AI tool. Before being approved for use, the tool underwent a Responsible AI review, informed by the University's AI Principles and governance processes. To support safe and responsible use, staff will be provided with tool-specific guardrails and guidance that outline appropriate use cases, key risks and practical steps to mitigate them.
The rollout will also be accompanied by Responsible AI and tool-specific training, resources and support to help you build the confidence and capability needed to use the technology effectively and responsibly. As with all enterprise technologies, the University will continue to monitor adoption, gather feedback and refine its approach over time.
When using enterprise AI tools, it's important to continue applying the same privacy, recordkeeping and data security obligations that apply to all University information.
Staff should only use personal information for authorised purposes and be mindful of privacy risks associated with AI, including offshore processing and globally routed inference. The quality of AI outputs depends on the quality of the information available, so outdated excessive or inaccurate data can lead to unreliable results.
Business areas can reduce risk by reviewing permissions, improving data classification and cleaning up legacy information. Staff should apply data minimisation principles, critically assess AI-generated outputs, and ensure transparency where AI contributes to decisions that affect individuals. Any use of personal information in AI tools needs to be based on strong principles regarding the fair and reasonable use of personal information.
Student AI learning will continue to be provided by the Teaching and Learning function.
The University will provide a range of training materials, guidance and support resources to help you develop practical AI skills in the new tool.
This includes ChatGPT Edu self-paced learning, online resources, knowledge articles, access AI communities of practice and other learning opportunities tailored to different needs and experience levels. More information will be provided on specific training available to you ahead of the Semester 1 2027 release.
The partnership with OpenAI includes teaching and research enablement, faculty support, student engagement and access to the OpenAI Academy.
Academic integrity remains a core priority for the University. Alongside access to AI tools, the University provides clear guidance on acceptable and responsible use, supports educators in designing assessments for an AI-enabled environment and helps students develop the skills needed to use AI responsibly and ethically. The focus is on fostering critical thinking, transparency, and responsible AI use, rather than avoiding the technology altogether.
For more information on academic integrity, refer to our Academic integrity page
For more information on using AI responsibly, familiarise yourself with our online guidance. It remains the responsibility of all students to use AI responsibly.
We're very aware the environmental impacts of AI are not yet well understood and could be significant — and we think that's precisely why a university should be leading on this, not standing back from it.
We're investing $1 million in an AI Sustainability Research and Innovation Fund to measure and reduce AI's environmental impacts — using our own multidisciplinary expertise in energy systems, cooling technologies and lifecycle assessment, which is already informing both the NSW and Federal parliamentary inquiries into data centres. A deployment of this scale gives us the opportunity, and the responsibility, to generate evidence the whole sector can use.
We're very aware the environmental impacts of AI are not yet well understood and could be significant, and are committed to exploring practical approaches to address these, alongside research opportunities to advance understanding and identify ways to develop and deploy AI that supports a more sustainable future.
Rather than investing in offsets, which are well intentioned but rely on immature and highly uncertain methodologies, we see this as a strategic opportunity to invest in research to understand and address the suite of environmental impacts of AI, employing the significant cross-disciplinary expertise of our academics. We're delighted to confirm our strategic $1 million seed funding for an AI Sustainability Research and Innovation Fund to support better understanding and measurement of the environmental impacts of AI, and ultimately develop the means to reduce them.
A working group with representatives across the University’s Sustainability, Digital Sydney and Research portfolios has been established to develop the Fund's governance, research priorities, operating model and implementation approach, alongside preparations for the roll out of ChatGPT Edu. We look forward to confirming details and outcomes in due course.
If you have a question that relates specifically to ChatGPT Edu, contact ict-chatgpt.edu@sydney.edu.au.
