CSIRO Industry PhD: Securing AI interactions - enhancing the MCP with governance, authentication, and real-time monitoring

This project strengthens the model context protocol (MCP), a framework designed to enforce contextual integrity in AI-driven interactions throughout their lifecycle, to improve AI security by addressing gaps in governance, authentication, and monitoring.

The outcome will be a robust security framework for MCP, enhancing AI system trustworthiness. This solution will improve AI security and reliability, benefiting industries and enabling safer AI deployment. 

Industry Engagement Component Description

- Deploy and test the “Securing AI Interactions” system within cloud-native environments.

- Participate in technical workshops and design sprints to align academic outputs with industry requirements.

- Access real-world datasets, cloud infrastructure, and secure lab environments.

- Regular mentoring sessions with mentor on secure system design, policy enforcement, and AI integration.

- Contribute to internal evaluations, product development feedback, and security reviews.

- Participate in industry events or conferences, to showcase project outcomes and network with cybersecurity professionals. 

The opportunity:

The CSIRO Industry PhD Program (iPhD) is a four-year research training program, focusing on applied research that benefits industry by solving real-world challenges. It aims to produce the next generation of innovation leaders with the skills to work at the interface of research and industry in Australia. 

The Program includes:

1. Admission to a university PhD program.
2. Supervision by the participating university, CSIRO, and an industry partner.
3. A four year scholarship package totalling approximately $47,000 per annum (2025 rate).  
4. A four year Project Expense and Development package of $13,000 per annum. 
5. A 60 day Industry Engagement component with the industry partner. 
6. A structured professional development and training program to develop your applied research skills.

Successful students are subject to the policies, procedures and guidelines of the participating university in addition to the CSIRO Industry PhD Program terms and conditions. Students will receive a standard PhD on completion.

Supervisory team details:

University

CSIRO

Industry

Essential skills:

• Background/qualifications in physics, engineering (electrical engineering, photovoltaic engineering)
• Background in machine learning and data analytics
• Understanding of renewable energy systems and smart grid technologies
• Knowledge of cybersecurity principles and data privacy frameworks

Desirable skills:

• Experience with security engineering, such as threat modelling, secure protocol design, identity management, or auditing mechanisms. 
• Hands-on experience with LLM alignment, evaluation, or red-teaming, especially in detecting anomalous or unsafe model behaviours.
• Ability to work across academia–industry collaboration, including requirements analysis, prototype development, and iterative system design.

Eligibility Requirements 

1. Be an Australian citizen or Permanent Resident, or a New Zealand citizen.  
2. Meet participating university PhD admission requirements. 
3. Meet university English language requirements. 
4. Not have previously completed a PhD.
5. Be able to commence the Program in the year of the offer. 
6. Enrol as a full-time PhD student. Part-time arrangements may be considered if approved by the supervisory team and in accordance with university policy.
7. Be prepared to be located at the project location(s) that the host university has approved and, if required, comply with the host university’s external enrolment procedures. 
8. Be prepared to undergo onboarding to CSIRO, which will include passing mandatory government background checks (allow for between 4 to 8 weeks) and complete any other CSIRO requirements. 

Application Process 

1. Applicants are required to submit an expression of interest (EOI), CV, transcripts and your motivation to the above supervisory team
2. The EOI is assessed by the supervisory team and shortlisted applicants are interviewed. 
3. The supervisory team nominates a preferred applicant and informs iPhD office and the Graduate Research School

Further information:

Please see the CSIRO website and the project site Securing AI interactions: enhancing the MCP with governance, authentication, and real-time monitoring or contact a member of the supervisory team above.