- The Australian Department of Home Affairs is establishing a new security platform to allow government agencies to use generative AI with sensitive data, classified from “OFFICIAL” up to “PROTECTED” level.
- In recent months, the Department has held “town hall” workshops with over 80 technology vendors, including Google Australia, AWS, IBM, Oracle, Microsoft, Macquarie Telecom, and Deloitte, to explain the new AI policy.
- The meetings focused on the AI guidance published in early October 2025, which defines principles for the safe use of generative AI within public record-keeping systems.
- Currently, only 18 AI vendors are permitted to access “OFFICIAL” data, including OpenAI (ChatGPT), Anthropic (Claude), and Google (Gemini) — these parties have passed security audits and achieved “Hosting Certification Framework” certification.
- The use of AI outside this list is still possible, but requires careful risk assessment regarding foreign ownership and influence (Foreign Ownership Control or Influence).
- The Department of Home Affairs’ AI policy was developed in collaboration with the Digital Transformation Agency (DTA) and the Australian Signals Directorate (ASD), aiming to standardize security benchmarks for AI implementation across the public sector.
- The Australian Department of Finance, led by Minister Katy Gallagher, recently announced a whole-of-government policy requiring all public servants to have access to safe generative AI directly on their laptops.
- All government agencies will be required to appoint a Chief AI Officer before July 2026 to oversee compliance with security policies and practical implementation.
- The Department of Home Affairs is preparing the next step: expanding the use of AI for “PROTECTED” level data, which demands stronger protective measures, before considering higher classifications like “SECRET” or “TOP SECRET.”
📌 Summary: The Australian Department of Home Affairs is establishing a new security platform to allow government agencies to use generative AI with sensitive data, classified from “OFFICIAL” up to “PROTECTED” level. Currently, only 18 AI vendors, including OpenAI (ChatGPT), Anthropic (Claude), and Google (Gemini), are permitted to access “OFFICIAL” data. The Department of Finance recently announced a policy requiring all public servants to have access to safe generative AI directly on their laptops. All government agencies must appoint a Chief AI Officer before July 2026 to oversee security policy compliance and practical implementation.
