The Joint Select Committee on Artificial Intelligence sought submissions from the public as part of its inquiry into a range of matters relating to artificial intelligence (AI) as per its Terms of Reference.
auDA's submission includes a whitepaper titled A national trust anchor for AI agents: Why the country-code Top Level Domain (ccTLD) regsitry should operate the national agent namespace.
You can read our summary below, or download the full whitepaper.
Dear Chair
I welcome the opportunity to make a submission to the Joint Select Committee on Artificial Intelligence.
In a future with millions or even billions of AI agents operating online, people and machines will need a safe and reliable way to know whether an agent is legitimate and who it represents. Today, a user visiting health.gov.au can have confidence that they are accessing an official Australian Government health service because the health.gov.au domain name provides a recognised and trusted link to that organisation. As governments and businesses increasingly deploy AI agents to interact with people and other machines, we will need a similar way to verify that an AI agent claiming to act on behalf of an organisation is genuinely authorised to do so. The Domain Name System (DNS) could provide that trust layer by linking AI agents to verified domain names and the organisations behind them.
I attach my whitepaper, A National Trust Anchor for AI Agents, for consideration by the Committee. The paper addresses an emerging aspect of artificial intelligence that I believe warrants consideration as Australia develops its national approach to AI: the identity, accountability and trust infrastructure required as AI systems evolve from tools that generate information into autonomous agents capable of acting across digital services and organisational boundaries. The paper is particularly relevant to paragraphs (k), (l) and (m) of the Committee's Terms of Reference: risks and harms arising from AI; national security and cyber security opportunities and risks, including the resilience of critical infrastructure; and the implications of emerging AI capabilities for Australia's national security and strategic resilience.
AI agents represent an important change in the nature of the internet. Increasingly, agents will be capable of discovering services, invoking tools, exchanging information, initiating transactions, making commitments and interacting with other agents with varying degrees of autonomy. As this occurs, many interactions that today involve human judgement will instead occur machine-to-machine and at machine speed.
This creates an emerging trust problem. Before accepting a consequential action from an AI agent, another system may need to establish independently what agent is acting, which person or organisation stands behind it, whether that relationship remains current, and where the authority claimed by the agent can be verified. These questions become particularly important in relation to fraud and impersonation, cyber security, government services, critical infrastructure, supply chains and national security.
The attached paper argues that accountability for AI agents should not depend solely on proprietary identity systems operated by the platforms on which those agents happen to run. Australia should consider establishing a platform-neutral, sovereign trust anchor through which an agent's durable identity and accountable sponsor can be independently resolved, while leaving authentication, credentials, authorisation and operational controls to the specialist systems designed to provide them.
The paper proposes that Australia's existing national DNS infrastructure warrants serious consideration as the foundation for such a trust model. The DNS already provides globally resolvable, hierarchical delegation under national stewardship, while DNS Security Extensions provide mechanisms for authenticating published information. The proposal is not that DNS should determine whether an agent is safe or authorised to undertake a particular transaction, but that it could provide the authoritative public starting point from which an agent's accountability and associated verification information can be established.
This also presents an opportunity for sovereign capability. Australia need not develop every foundation model or AI platform domestically to retain sovereign control over important elements of the trust infrastructure through which autonomous agents interact with Australian organisations, citizens, government and critical systems.
The agent ecosystem is still developing, and the paper does not suggest that Australia should prematurely prescribe a particular technical implementation. Rather, I submit that Australia should recognise machine-verifiable agent identity and accountability as an emerging component of national digital infrastructure, and begin the technical, policy and multistakeholder work required to determine how that capability should be provided.
In particular, I invite the Committee to consider recommending that the Australian Government:
- Recognise trusted, independently verifiable identity and accountability for AI agents as an emerging national cyber security and digital infrastructure requirement;
- Convene government, cyber security, digital identity, AI safety, critical infrastructure, privacy, competition and Internet-governance stakeholders to examine requirements for a sovereign, platform-neutral trust anchor for AI agents; and
- Support controlled technical trials and international standards engagement to test interoperable approaches before proprietary agent identity systems become entrenched.
These measures would be proportionate to the maturity of the technology. They would not require Australia to determine the final architecture today but would ensure that the country is actively shaping the trust infrastructure on which safe agentic AI may increasingly depend.
Thank you for considering this submission. auDA would welcome the opportunity to discuss these findings and provide further information. Please contact me at glenn.wightwick@auda.org.au.
Yours sincerely
Glenn Wightwick
Chief Innovation Officer
.au Domain Administration (auDA)