What 497 submitters told the committee
The pipeline sorted the submitters into 37 topics and 4 themes, and wrote a claim and a description for each. This page shows those claims with the figures behind them. Every number is computed when the page loads.
- 01National AI capabilityturn infrastructure and adoption into lasting Australian capacity under enforceable safeguards22 topics · 435 submitters
- 02AI content use and abusePut preventive duties on businesses best able to act8 topics · 139 submitters
- 03AI at workWorker voice and safeguards should shape job redesign5 topics · 51 submitters
- 04AI educationBuild human capability through coordinated, teaching-led reform2 topics · 23 submitters
- Supports
- Builds on
- Mixed
- Redirects
- Opposes
- Unclear
National AI capability
turn infrastructure and adoption into lasting Australian capacity under enforceable safeguards
The sources argue that Australia should move beyond pilots and passive dependence on overseas providers by building the infrastructure, skills, institutions and implementation capacity needed to create lasting national value. They also support safeguards based on a system’s real-world consequences and authority, including human accountability and independent assurance for high-stakes uses. They differ on whether priority should go to public procurement, compute and data centres, local firms and robotics, sector adoption and education, or stronger controls for autonomous, health and national-security systems.
- Positions
- 85% supportive · 115 contesting · 69 unclear
- Sector
- Individual 35% · Company 24% · Civil Society 15% +3
- Most-cited
- +189
Topics in this theme 22
One framework must span all harms
AI policy should pursue public value and practical national capability through enforceable obligations scaled to capability, autonomy and consequence. Systems should remain auditable, evidence-bound and subject to human authority, continuous assurance, incident visibility and effective intervention. Unlike clusters separating frontier risk from present-day regulation, this proposition deliberately places catastrophic scenarios and immediate harms within one governance framework.
309 submittersInvest for control and continuity
This cluster defines sovereign AI as practical control, resilience and strategic choice rather than total self-sufficiency or purely domestic hosting. It calls for targeted investment in defensible Australian applications, people, institutions and implementation expertise, using procurement and public funding to move proven capabilities from pilots into sustained production. Unlike the research-and-skills sovereign cluster, it places greater emphasis on reducing concentrated foreign dependence and converting existing Australian capability into durable operational capacity.
210 submittersGovernment purchasing should build enduring capability
AI policy should pursue lasting national and public value by targeting areas where Australia can retain capability, intellectual property and export returns. Adoption policy should move beyond pilots to sustained workflow redesign, with government acting as a responsible lead adopter, capable buyer and market maker. Unlike other sovereignty clusters, this proposition makes public procurement and exemplary government adoption the primary catalysts for capability.
131 submittersConsequences and authority should determine duties
AI regulation should scale with a system’s consequences, delegated authority and operating context, building on existing law while adding binding duties where gaps remain. Consequential decisions should be transparent, contestable and subject to meaningful human authority, with evidence sufficient to reconstruct actions and changes. Unlike agent-only governance clusters, this framework applies generally but imposes stronger runtime controls when systems can act autonomously.
105 submittersConvert growth into national value
This cluster argues that data centres should be developed as strategic national capability rather than treated mainly as property investments. It seeks consistent enforceable standards, coordinated approvals, additional clean energy, efficient resource use, infrastructure resilience, place-based consultation and community benefit-sharing so growth produces lasting Australian economic and technological value. Unlike the broader emerging-framework cluster, it advances firmer development obligations and gives particular weight to community consent and domestic value capture.
95 submittersEmpower independent assurance for autonomous AI
Advanced and autonomous AI should be treated as a national-security and strategic-resilience issue requiring independent, capability-based assurance. Enforceable operational controls, human accountability, serious-incident reporting and shared learning should be backed by an Australian AI Safety Institute with sufficient authority, access and technical capacity. Unlike general assurance clusters, this proposition focuses on high-consequence systems and a specifically empowered national institution.
78 submittersPrioritise physical AI and scale-ups
Australia should build practical AI capability while remaining connected to global technology, concentrating investment where local knowledge, data and operating conditions create an advantage. Policy should convert research and pilots into customers, scalable firms and exports, give physical AI and robotics a dedicated pathway, and support regional adopters. Unlike general sovereignty clusters, this cluster foregrounds commercial scale-up and embodied AI industries.
45 submittersHosting deals must secure national benefits
AI computing capacity should be treated as strategic national infrastructure, with sovereignty measured by practical control and access rather than physical location alone. Hosting agreements and public support should secure Australian benefits while building domestic capability alongside foreign investment and independent safety expertise. Unlike clusters focused on restricting data centres or reducing provider dependence, this cluster centres on structuring compute-hosting arrangements to capture strategic value.
40 submittersGovern health data nationally
This cluster proposes treating healthcare AI as a coordinated national capability and health data as strategic infrastructure subject to trusted governance. It calls for privacy, consent, community control, risk- and use-based clinical regulation, real-world performance monitoring and retained human clinical responsibility. Unlike general AI assurance clusters, it addresses the distinctive ethical, data-governance and implementation requirements of clinical settings and makes public trust central to health-sector deployment.
34 submittersLayered capability must preserve strategic access
Sovereign AI should be defined by practical control and resilience against foreign providers that may withdraw, restrict, reprice or disrupt access. Australia should build computing and frontier-training capacity where strategically useful, attach national-interest conditions to major infrastructure and develop a layered ecosystem rather than a single national model. Unlike broader domestic-ecosystem proposals, this cluster is specifically organised around concentrated-provider risk and continuity of access.
34 submittersPause expansion pending public-interest proof
Rapid data-centre expansion should pause or face restriction until proponents demonstrate sustainability and public benefit. Assessment should address cumulative pressure on water, ecosystems, climate resilience, energy systems, household costs, local amenity and vulnerable communities, with early participation and disclosure of influential interests. Unlike infrastructure-planning clusters that support conditional development, this cluster adopts a precautionary presumption against further expansion.
34 submitters1 SectorFund the full integration pathway
This cluster contends that productive integration—not superficial uptake—should be the objective of small-business AI policy. It seeks practical assistance across the entire adoption pathway, delivered in short, sector-specific forms by trusted intermediaries, with workforce development focused on work redesign, judgement and continuing participation. Unlike the robotics cluster, it addresses AI adoption broadly and identifies limited organisational readiness and implementation capacity, rather than technology cost alone, as the principal barriers.
33 submittersTrusted support should embed productive use
Policy should shift from encouraging initial uptake to helping organisations embed AI productively in workflows. Trusted implementation support should address uneven practical constraints, while proportionate safeguards and clearer sector-specific laws preserve professional and human accountability. Unlike the small-business cluster, this proposition applies across sectors and particularly emphasises legal clarity for professional use.
29 submittersBuild selective supply-chain capability
Sovereign AI should mean practical control, resilience and choice rather than complete self-sufficiency. Australia should invest selectively across high-value parts of the supply chain, including domestic training and trusted early access, while strengthening testing, accountability, skills and research. Unlike sovereignty clusters centred on infrastructure approvals or strategic assets, this cluster focuses on creating a durable, interconnected domestic AI ecosystem.
23 submittersEnforce limits where actions occur
Agentic AI should be governed according to the authority and access it receives, with identity, permissions and operating limits enforced at the point where an action is attempted. Oversight should be continuous, independently testable and responsive to change, while consequential decisions remain subject to human and institutional accountability. Unlike transaction-law and trust-anchor clusters, this cluster centres on runtime control of agent behaviour.
18 submittersScale implementation with proportionate safeguards
Australia should accelerate practical AI adoption across the economy to improve productivity and services, closing the gap between experimentation and organisation-wide implementation, especially for small and medium enterprises. Skills, infrastructure, investment and pathways for Australian providers should be paired with clear safeguards that preserve innovation and protective AI uses. Unlike small-business-only clusters, this proposition presents a broad national adoption agenda spanning capability, security and liability.
12 submittersEvaluate deployed systems and changes
This cluster argues that assurance for consequential AI must rest on evidence from the system as actually deployed, with agent actions, permissions and outcomes capable of reconstruction. It calls for version-linked approval, reassessment after material changes, common evidence standards, independent Australian evaluation capacity and preserved rights of human review and challenge. Unlike broader governance clusters, it specifies a technically grounded assurance model centred on operational evidence, traceability and change control.
10 submittersVerify authority under existing laws
This cluster argues that transacting AI agents require verifiable authority, bounded permissions, evidence of actions and accessible recourse, but that existing laws should be tested before new AI-specific duties are imposed. It seeks coordinated regulation, market monitoring, shared governance infrastructure and technical standards alongside commercial identity and assurance services. Unlike the sovereign trust-anchor cluster, it focuses on the practical legal and operational controls surrounding particular transactions rather than creating a central public trust institution.
7 submittersMake councils national delivery partners
This cluster proposes making local government a formal partner in national AI policy and delivery rather than a downstream implementer. It seeks shared capability for smaller and regional councils, nationally consistent public-sector safeguards, responsible supplier requirements and workforce investment that protects professional judgement and careers. Unlike general public-sector or small-business adoption clusters, it focuses on closing capability gaps between councils and using AI to redesign local services rather than merely automate existing processes.
5 submittersIndependent capacity should balance growth risks
A national AI strategy should capture economic benefits, protect people and distribute gains broadly through adaptable risk-based regulation grounded in existing law. Australia should build independent technical capacity to test systems and monitor change while supporting adoption through skills, infrastructure and domestic capability. Unlike the institutional-resilience framework, this cluster explicitly integrates national infrastructure requirements with economic, community, security and sovereignty considerations.
5 submittersCoordinate pauses until safeguards work
Governments should slow or pause frontier AI development until enforceable safety thresholds, independent oversight and effective control methods are in place. The case rests on potentially catastrophic harms, the inadequacy of voluntary company restraint and the danger that competitive or military incentives will outrun governance. Unlike risk-management clusters that permit continued development, this cluster advocates internationally coordinated pacing or restriction.
4 submittersEstablish a sovereign neutral anchor
This cluster proposes a sovereign, platform-neutral trust anchor through which AI agents can establish authority and accountability without concentrating transaction powers or sensitive data in one system. It calls for an explicit public mandate, oversight, competition protections, federated assurance and immediate standards trials without prematurely fixing the final architecture. Unlike the cluster governing individual agent transactions under existing law, this cluster is principally concerned with the institutional trust layer underpinning the wider agent ecosystem.
2 submitters
Read the themekey points and quotes · who says what · submitters that push back
AI content use and abuse
Put preventive duties on businesses best able to act
The sources argue that creators, cultural custodians, abuse survivors and scam victims should not carry the main burden of detecting unlawful or harmful AI practices. They call for developers, platforms and other commercial actors to prove lawful content use, prevent foreseeable harms and provide practical redress. The sources differ on the tools required, including licensing and payment, Indigenous-led protections, procurement rules, safety duties, abuse detection and shared responsibility across the fraud chain; some also distinguish commercial training from legitimate public-interest uses.
- Positions
- 82% supportive · 23 contesting · 26 unclear
- Sector
- Industry Group 24% · Civil Society 22% · Individual 21% +3
- Most-cited
- +44
Topics in this theme 8
License commercial training, preserve public interest
Commercial AI training should rest on consent, licensing and remuneration, while policy should distinguish it from legitimate public-interest uses rather than impose a blanket rule. Providers should keep useful records of training data, provenance and legal authority, with affordable enforcement applying to offshore services in Australia. Unlike other copyright clusters, this proposition is distinguished by its explicit commercial-versus-public-interest boundary.
97 submittersShift responsibility to capable providers
This cluster treats AI-enabled harm as a systemic product and service risk rather than merely the result of individual misuse. It calls for enforceable, lifecycle-wide Safety by Design duties placed primarily on providers and other institutions best able to prevent harm, with priority given to children, gender-based violence and disproportionately affected groups. Unlike the survivor-centred abuse cluster, it supplies a broader responsibility model for closing legal gaps through technology-neutral, harm-based regulation.
52 submittersIndigenous-led protection must govern AI use
Commercial AI should use protected material lawfully and transparently through workable licensing and compensation systems accessible to smaller rights holders. Distinct Indigenous-led protection is required for cultural knowledge and expression, alongside rules for ownership, attribution, liability, digital replicas and practical redress. Unlike general copyright clusters, this cluster combines creator rights with culturally specific authority and protection.
45 submittersProcurement should reward provably lawful AI
Consent, creator control and payment should underpin AI content use, with transparent and traceable records demonstrating lawful access. Scalable licensing must address bargaining inequality, enforcement barriers and economic displacement while accommodating disagreement over collective models. Unlike clusters prescribing a single licensing structure, this proposition emphasises government procurement, funding and market access as incentives for lawful conduct.
35 submittersDevelopers must prove licensed content use
Developers should obtain consent and licences for distinct uses of protected content, make those uses traceable and bear the burden of demonstrating compliance. Inclusive compensation mechanisms, procurement standards, provenance infrastructure and source visibility should support creators and the sustainability of news. Unlike voluntary-market proposals, this cluster calls for stronger developer obligations and compensation systems that do not depend solely on private deals.
18 submittersImpose survivor-centred preventive duties
This cluster argues that gender-based violence, child exploitation and harms to vulnerable people are foreseeable AI safety risks requiring enforceable preventive duties on developers, deployers and commercial beneficiaries. It seeks binding, technology-neutral and risk-based laws, equality assessments, platform detection of abuse patterns, transparency and rapid survivor-centred remedies. Unlike the broader systemic-harm cluster, it concentrates on consequential abuse affecting vulnerable groups and specifies equality analysis and survivor-focused redress.
14 submittersRegulate the entire fraud chain
Because AI makes established scams faster, cheaper and more scalable, the Scams Prevention Framework should cover every organisation involved in creating, distributing and monetising fraud. Prevention, verification, platform duties and rapid remedies should replace overreliance on consumer vigilance, with losses shared more fairly among preventable-harm actors. Unlike broader online-safety clusters, this cluster targets synthetic identity, document and voice fraud across financial and public services.
8 submittersNegotiated voluntary markets should remain primary
Voluntary, negotiated licensing should remain the main route by which AI developers access copyright material, using flexible arrangements rather than a compulsory or uniform scheme. Government intervention should depend on demonstrated market failure, while licensing infrastructure should improve consent, rights clearance, payment and creator visibility. Unlike clusters seeking mandatory collective compensation, this cluster prioritises market negotiation and individual creator control.
6 submitters
Read the themekey points and quotes · who says what · submitters that push back
AI at work
Worker voice and safeguards should shape job redesign
The sources argue that workplace AI should improve job quality and share benefits, with workers involved in adoption and supported through training, redeployment and enforceable protections. Some focus on employees’ rights to transparency, privacy, explanation and human review, while others call for nationally consistent transition planning and stronger safeguards in essential or safety-critical work. They also differ on predicted job losses, with some expecting AI to change tasks and skills more often than eliminate whole occupations.
- Positions
- 92% supportive · 4 contesting · 2 unclear
- Sector
- Civil Society 49% · Industry Group 13% · Individual 13% +3
- Most-cited
- +12
Topics in this theme 5
Worker voice must govern adoption
Workplace AI should be judged by job quality and shared public value, with workers consulted early, meaningfully and continuously about adoption. Employees should receive transparency, privacy, explanation, challenge and accountable human decision-making, alongside training, redeployment and displacement protection. Unlike the broader labour-transition cluster, this cluster details individual and collective procedural rights within AI-enabled workplaces.
33 submittersRedesign work before predicting displacement
This cluster contends that AI will more often alter tasks and skill requirements than eliminate whole occupations, and that productivity depends on investment in people, management and work design rather than technology alone. It calls for broad continuing AI capability, transition support for unequally exposed groups, worker involvement and retained human accountability. Unlike trade-specific or small-business skills clusters, it addresses economy-wide job transformation and rejects simplistic occupation-loss forecasts.
25 submittersEnforce national safeguards and shared benefits
Worker participation and a fair share of AI’s benefits should be central to adoption, supported by nationally consistent and enforceable workplace safeguards. Workforce impact assessment, skills, redeployment and stronger protections should address displacement, degraded job quality and algorithmic management, with added safeguards in safety-critical and essential services. Unlike the workplace-rights cluster, this proposition emphasises national consistency, transition planning and heightened sectoral protections.
14 submitters1 SectorExpand capacity beyond current forecasts
This cluster argues that AI infrastructure is fundamentally a workforce challenge because data centres will intensify competition for electricians already needed in energy, housing and other major projects. It seeks expanded apprenticeships, stronger vocational and post-trade development, and support for employers facing cost, supervision and training-capacity constraints. Unlike broader AI-skills clusters, it concentrates on electrical trades and warns that existing workforce forecasts may materially understate demand.
4 submittersWithholding should fund protected employment
Clients purchasing overseas professional services should withhold contributions that support protected Australian employment and offset expected displacement costs. Rates should differ by service category and provider registration, reflect where workers performing the service reside, and be reviewed regularly with clear public guidance. This cluster is distinct from general AI labour-market proposals because it advances a specific fiscal mechanism targeting offshore professional-service substitution.
1 submitter1 Sector
Read the themekey points and quotes · who says what · submitters that push back
AI education
Build human capability through coordinated, teaching-led reform
The sources argue that Australia needs a coordinated national approach to AI literacy and education that strengthens human capability rather than promoting AI use for its own sake. They call for sustained support for educators, fair access and assessment methods that verify learning without blanket bans or unreliable detection tools. One focuses on teaching goals, age-appropriate use and standards for educational products, while the other focuses on funding, delivery through institutions and assessment reform.
- Positions
- 97% supportive · 0 contesting · 1 unclear
- Sector
- Civil Society 29% · Individual 29% +4
Topics in this theme 2
Fund educators and redesign assessment
Australia should turn existing education and skills frameworks into a coordinated, funded pathway for AI capability across education and working life. Investment should flow through educators and institutions, with assessment redesigned to verify learning without blanket bans or unreliable detection tools and with equity built into access and evaluation. Unlike the education-policy cluster, this cluster concentrates on implementation machinery and assessment reform.
21 submittersTeaching-led AI should build human capability
Education policy should make human capability its objective rather than treating AI uptake as an end in itself. A coordinated national plan should support age-appropriate, teaching-led use, comprehensive AI literacy, sustained educator development and assurance standards for educational products. Unlike the education-delivery cluster, this proposition primarily defines the pedagogical purpose and boundaries of responsible AI use.
10 submitters