AI Inquiry Atlas

An unofficial, independent guide to the public records of the Joint Select Committee on Artificial Intelligence. Not connected with the Parliament of Australia or the committee. Its makers also made a submission (no. 481; see About). Summaries and scores are written by AI and can be wrong: check the original before relying on them.

Submission 38

Raedan AI

Submitter group: Companies.

An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.

Open in the interactive guide

Summary

Read the original submission for the full text.

Raedan AI, a small Australian company working in semantic AI, argues that government AI governance should add a whole-of-government 'semantic control' framework for authoritative meaning, context and evidence, delivered through federated standards preserving agency authority. It makes seven recommendations on architecture, assurance, accountability, interoperability, sovereign infrastructure and a demonstration program.

Key points

Quotes are the submitter’s own words. Text taken from documents, including scanned pages read by text-recognition software, can differ from the PDF. Check the original. Listed in the same order as on the submission’s page in the interactive guide.

  1. Argues that government governs data access and AI model use but lacks a whole-of-government mechanism for governing authoritative meaning across AI systems.

    The Australian Government already governs access to data and use of AI models. It does not yet have a whole-of-government mechanism for governing authoritative meaning across AI systems.

    A program corrected obvious text-recognition errors in this quote for display.

  2. Argues that agents across agencies may use the same term with different definitions, creating semantic fragmentation as agentic AI scales.

    Two agents operating across agencies might use the same word while relying on different definitions, policy versions or authority sources. At scale, this creates semantic fragmentation: many systems encode different interpretations of government concepts and pass those interpretations to other systems.

  3. Says the issue is institutional: government must decide which definitions and rules are authoritative and who owns them.

    The resulting issue is institutional rather than purely technical. Government needs a method for deciding which definitions, rules, relationships and evidence are authoritative, who owns them, when they apply, how changes are versioned and how AI systems consume them.

    A program corrected obvious text-recognition errors in this quote for display.

  4. Proposes a governed framework in which shared standards coexist with agency and domain authority, rather than one central ontology for the Commonwealth.

    The policy objective is not one central ontology for the Commonwealth. The objective is a governed framework in which shared standards coexist with agency and domain authority, while AI systems receive traceable institutional context.

    A program corrected obvious text-recognition errors in this quote for display.

  5. Argues that sovereign AI should include control over government language, definitions, rules and domain knowledge.

    Sovereign AI is not only sovereign compute, models and data. It also requires sovereign control over government language, definitions, rules and domain knowledge.

    A program corrected obvious text-recognition errors in this quote for display.

Recommendations

Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.

  1. Establish a whole-of-government Semantic Control Framework with principles and minimum controls for authoritative meaning, context, relationships, rules, evidence, provenance, ownership, effective dates and permitted use in AI.

    Establish a Whole-of-Government Semantic Control Framework. Define principles and minimum controls for authoritative meaning, context, relationships, rules, evidence, provenance, ownership, effective dates and permitted use in AI.

    A program corrected obvious text-recognition errors in this quote for display.

  2. Include semantic capabilities as reusable infrastructure within the Australian Government Architecture, with hybrid and neuro-symbolic reference patterns for policy-intensive use cases where appropriate.

    Recognise semantic capabilities as reusable government infrastructure within the Australian Government Architecture alongside data, integration, identity, security and AI services. Include hybrid and neuro-symbolic reference patterns for policy-intensive use cases where appropriate.

    A program corrected obvious text-recognition errors in this quote for display.

  3. Extend AI assurance to semantic assurance: for high-impact AI, assess interpretation against authoritative definitions and require traceability between model inference, source evidence and governed semantic assets.

    Extend AI assurance to semantic assurance. For high-impact AI, assess whether key concepts are interpreted against authoritative definitions, rules and context. Require traceability between model inference, source evidence and governed semantic assets.

    A program corrected obvious text-recognition errors in this quote for display.

  4. Clarify accountability for meaning across information owners, Chief Data Officers, Chief AI Officers, CIOs, AI Accountable Officials, information governance practitioners and enterprise architects, preserving agency and domain authority.

    Establish clear accountability for meaning. Clarify responsibilities across information owners, Chief Data Officers, Chief AI Officers, Chief Information Officers, AI Accountable Officials, information governance practitioners and enterprise architects. Preserve agency and domain authority.

    A program corrected obvious text-recognition errors in this quote for display.

  5. Develop semantic interoperability standards: machine-readable exchange patterns for definitions, context, relationships, provenance and authority across agencies, platforms and AI agents, avoiding a mandatory single government-wide ontology.

    Develop semantic interoperability standards. Define machine-readable exchange patterns for definitions, context, relationships, provenance and authority across agencies, platforms and AI agents. Avoid a mandatory single government- wide ontology.

    A program corrected obvious text-recognition errors in this quote for display.

  6. Include semantic assets (government terminology, legislative concepts, domain vocabularies, knowledge models, business rules) within sovereign AI planning, alongside compute, models and data.

    Include government terminology, legislative concepts, domain vocabularies, knowledge models and business rules within sovereign AI planning, alongside compute, models and data.

    A program corrected obvious text-recognition errors in this quote for display.

  7. Establish cross-agency semantic-control demonstrations, including a policy-intensive domain where AI interprets unstructured information against explicit legislation, policy rules and authoritative sources.

    Establish a whole-of-government demonstration program. Evaluate semantic control in several cross-agency use cases, including at least one policy-intensive domain where AI interprets unstructured information against explicit legislation, policy rules and authoritative sources.

    A program corrected obvious text-recognition errors in this quote for display.

Read the original

In some quotes a program corrected obvious text-recognition errors for display.

Open in the interactive guide