As universities worldwide grapple with how to harness generative AI without compromising privacy, integrity or trust, Macquarie University has taken a decisive step, building and rolling out a secure, in-house AI platform to its entire workforce.
The ambitious initiative positions Macquarie as one of the first Australian universities to deploy enterprise-grade generative AI at scale – ethically, securely and by design – while providing all 3500 staff with access to leading AI models in a trusted environment.
The custom-built platform, ChatMQ, has been designed and developed in-house by Macquarie’s own team and operates under the University’s Responsible AI Policy. It provides staff with secure access to multiple leading generative AI models, including OpenAI’s GPT-4o and Anthropic’s Claude Sonnet 4.5, all within an Australian-based environment aligned with the University’s governance and values.
Rather than acting as a consumer-style chatbot, ChatMQ functions as a “trusted AI workspace”, empowering staff to confidently use generative AI with university documents and data they would normally be prohibited from uploading to public platforms. This capability is already transforming how staff approach teaching design, research support and operational workflows.
“This is about giving our people the confidence to use AI responsibly, not asking them to work around its risks,” Deputy Vice-Chancellor (People and Operations) Professor Eric Knight said.
“By investing in secure, enterprise-grade AI that aligns with our values, we’re strengthening Macquarie’s competitiveness and demonstrating how AI can empower research and education while protecting academic integrity and privacy.
“As higher education institutions wrestle with the risks and rewards of generative AI, Macquarie’s approach signals a shift toward institutional ownership – treating AI not as a consumer add-on, but as core infrastructure governed by clear ethical and operational standards.”
Rather than purchasing an off-the-shelf solution, Macquarie adopted a deliberate co-development approach, shaping ChatMQ through an alpha phase with 100 staff followed by a 400-person beta group. Feedback from across faculties and professional services was incorporated at every stage, ensuring the platform reflected real-world needs and workflows.
That approach has paid off. Among early users, 89 per cent said ChatMQ met or exceeded expectations, while 85 per cent rated the support and training as highly useful.
Unlike consumer AI platforms, ChatMQ does not train large language models on user inputs and processes data within Australia under its current design – unlocking practical, high-value use cases that were previously off-limits.
“Knowing I wasn’t breaching privacy made all the difference,” one staff member said. “I could finally work with real MQ documents without risk.”
Macquarie’s Head of AI, Phil Laufenberg, said the decision to build in house reflected a broader shift in how institutions are responding to generative AI.
“AI is reshaping universities whether we act or not,” Mr Laufenberg said. “We chose to take ownership, building a governed, multi-model platform that empowers our community while embedding ethics, safety and human oversight from day one.
“And as Australia's first purpose-built modular AI solution built using open-source libraries, ChatMQ also represents a significant step forward for Australia in the development of sovereign AI capability.
“Our hope is that ChatMQ demonstrates how institutions can adopt generative AI at scale – confidently, responsibly and on their own terms.”