Submission 282
Australian Council of Trade Unions (ACTU)
Submitter group: Unions.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
Argues workers must shape AI adoption and share its benefits. Recommends workplace rights, targeted frontier-model licensing, enforceable data-centre standards, copyright protections and a Digital Duty of Care. Proposes government-coordinated pro-worker sovereign AI without training a frontier model; a joint researcher-union supplement details a coordination vehicle, Australian-controlled compute, transferable training and public-value safeguards.
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.
Argues that appropriate worker voice is a precondition for real and sustained productivity gains from AI.
Appropriate worker voice is a precondition to achieving any real and sustained productivity gains.
Argues that existing workplace laws were not built for algorithmic management and automated decision-making at current scale.
These frameworks were not built for algorithmic management and delegated automated decision-making at the scale now occurring, and the gaps are widening as adoption accelerates. Nearly the entirety of the Fair Work Act and all of our WHS laws and regulations were established prior to the advent of modern generative AI and its proliferation through workplaces.
Proposes a government-coordinated national initiative to build pro-worker sovereign AI, bringing together government, unions, universities, industry and superannuation to enhance control over AI infrastructure.
To achieve this, Australian Unions are calling for a government-coordinated national initiative to build Pro-Worker Sovereign AI in Australia. This will protect the fair go in the digital age, bringing together government, unions, universities and industry experts, the superannuation sector, and the Australian tech industry to enhance our control over AI infrastructure.
Proposes that the AI Safety Institute oversee a licensing regime for large frontier and open-weight models.
Principally, by empowering the Artificial Intelligence Safety Institute (AISI) to oversee a licensing regime for the development and deployment of large frontier models and large open weight models in Australia to keep Australians safe.
Argues that traditional copyright is insufficient to address AI's harms to creative and media workers and that a different approach is needed.
However, traditional copyright is insufficient to address the harms that artificial intelligence poses to creative and media workers in the form of undercutting, replacement, and substitution. A different approach is needed.
Argues that the Digital Duty of Care should impose clear, enforceable obligations on AI platforms rather than leaving safety processes to providers' discretion.
However, the Duty will only be effective if it imposes clear and enforceable obligations on companies. Unions believe service providers should not be left with excessive discretion to determine for themselves what constitutes an adequate safety process.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Empower the AI Safety Institute to oversee a licensing regime for the development and deployment of large frontier models and large open-weight models in Australia.
Principally, by empowering the Artificial Intelligence Safety Institute (AISI) to oversee a licensing regime for the development and deployment of large frontier models and large open weight models in Australia to keep Australians safe.
Consider the 'Copyright Plus' proposal so that all creators are fairly recognised and paid for their work, and take action to protect First Nations Cultural and Intellectual Property.
Action to protect creators and their works from theft by large AI companies without consent or remuneration, including consideration of the ‘Copyright Plus’ proposal to ensure that all creators are fairly recognised and paid for their work. This includes particular action to protect First Nations Cultural and Intellectual Property.
Require licensing to provide ongoing transparency, regular safety auditing, high-risk-use guardrails, effective governance and mandatory major-incident reporting.
A licensing regime would mandate ongoing transparency over AI products and models available in Australia, including regular auditing of their safety, the development of guardrails against high- risk use, effective governance, and mandatory reporting of major incidents.
Develop human-focused deployment models before introducing automated vehicles, addressing negative transport-workforce impacts.
which details how reforms to laws concerning the effects of driverless vehicles are proceeding with no corresponding proposal to assess, prevent, or regulate negative impacts on the transport workforce. We refer the Committee to the TWU’s submission for further important details in this example and the need for human-focused deployment models to be developed before automated vehicles are introduced.
Require consultation at the decision to adopt AI, rather than waiting for redundancies.
In practice, employers are ignoring their obligation to consult on the decision to introduce AI but rather are consulting when redundancies take place. The decision which should trigger consultation is the decision to adopt artificial intelligence.
Require enforceable consultation before AI-adoption decisions are finalised, with information, time and capacity for workers and unions to influence outcomes, employer mitigation of adverse outcomes and effective remedies.
An enforceable 'Introduction of Change' regime requiring genuine consultation before decisions to introduce AI are finalised, with workers and their unions given the information, time and capacity to influence outcomes, obligations on employers to mitigate adverse outcomes identified during consultation, and effective remedies for non-compliance.
Increase redundancy pay for workers if evidence shows increasing numbers of workers are displaced or roles automated by AI.
An increase in redundancy pay for workers if evidence shows increasing numbers of workers are displaced or roles automated by AI.
Guarantee workers access to digital and AI tools, with corresponding policies and procedures on how and when to use AI at work.
Workers must have guaranteed access to digital and AI tools, with corresponding policies and procedures for how and when to use AI at work.
Have the Commonwealth establish and run a standardised, free, universally accessible AI training course (covering the basics, functionality, ethics and AI shortfalls), with paid time off for workers to attend.
The Commonwealth Government should establish and run a standardised, free and universally accessible AI training course. Employers should provide paid time off for all workers to attend training. The course should include: • The basics; what AI is, where data comes from and how it is stored, • Functionality; AI capabilities and how it can be used in different settings, with some examples from various widely used models like Chat GPT, Co-Pilot and Clause, • Ethics; when is it appropriate to use AI and how to protect clients and other vulnerable people, and • AI shortfalls; including hallucinations and mis/disinformation.
Provide notification of AI-made or substantially assisted worker decisions, meaningful explanations and appeal to an authorised human decision-maker; require substantive oversight with review time included in workloads and staffing.
Workplace ADM transparency and human oversight rights: a right to know when AI or ADM is used to make or substantially assist decisions affecting workers, a right to a meaningful explanation, and a right of appeal to a human decision-maker with genuine authority. Human oversight must be substantive, with the time required to properly review ADM outputs factored into workloads and staffing.
Remove the employee-records exemption and enact workplace surveillance and worker-data protections covering collection, use and retention, including AI training and performance standards.
Removal of the employee records exemption from the Privacy Act and the enactment of dedicated workplace surveillance and worker data protections, governing the collection, use and retention of worker data, including its use to train AI systems and to set or measure performance standards.
Give workers an affirmative right to refuse to use technology that is unsafe or in conflict with their professional or registration obligations, without detriment.
An affirmative right for workers to refuse to use technology that is unsafe or in conflict with their professional or registration obligations, without detriment.
Prohibit undisclosed AI-set performance standards: workers to be told benchmarks, tolerances and triggers, with no discipline on AI-generated metrics without human verification and procedural fairness.
Prohibition of undisclosed AI-set performance standards: workers must be told the benchmarks, tolerances and triggers against which they are measured, and no disciplinary action should proceed on AI-generated metrics without human verification and procedural fairness.
Clarify that liability for AI tool errors lies with deploying employers, with no honest and reasonable mistake defence for systemic AI-facilitated breaches, and review to close loopholes.
Government firmly clarifies liability for AI tool errors lies with the employers that deploy the tools/systems/applications, with no honest and reasonable mistake defence available for systemic breaches facilitated by AI. Review to identify and close any loopholes.
Impose enforceable obligations on employers adopting AI to invest in skills, training and transition support, including redeployment and retraining. so affected workers share adoption gains.
Enforceable obligations on employers adopting AI to invest in skills, training and transition support, including redeployment and retraining, so the gains of adoption are shared with the workers affected.
End unremunerated creative-work ingestion, protect Indigenous Cultural and Intellectual Property by statute, and invest publicly in sovereign AI aligned with Australian values, a stable tax base and economic resilience.
Action to end the unremunerated ingestion of creative work, statutory protection of Aboriginal and Torres Strait Islander Cultural and Intellectual Property, and public investment in sovereign AI capability so that the technology is developed in line with Australian values, with a stable tax base and economic resilience.
Consider establishing a specialist Expert Panel on AI within the Fair Work Commission to set standards on AI in workplace relations, such as rostering, surveillance, performance management, hiring and termination.
Consideration of a specialist Expert Panel on AI to be established within the Fair Work Commission, tasked with setting standards regarding the use of AI in workplace relations, including for example: rostering, surveillance, performance management, hiring and termination
Coordinate a government-led national initiative to build Pro-Worker Sovereign AI, aligned with national economic priorities and tied to secure, safe and well-paid jobs.
We are asking for government to provide coordination of a national endeavour to build AI that boosts national priorities (such as Future Made in Australia) with the explicit purpose of protecting and creating secure, safe, and well-paid jobs.
Coordinate a tripartite industrial planning and investment vehicle across the AI stack, seeding multiple partnerships with universities, technology firms, businesses and investors.
The Australian Government should coordinate an industrial planning and investment vehicle to build Australia’s sovereign AI capability where possible across the tech stack and seed the development of other AI projects in collaboration and partnership with universities, the tech sector, businesses and investors.
Develop smaller models suited to Australian economic and industrial purposes using Australian datasets, while protecting creative works.
Rather, critical capability could look like models that are smaller than the leading frontier models but are designed for specific cultural and economic needs, and utilising Australian datasets and protecting critical Australian data. Unions believe we should focus on building models of a size and scale that makes sense in Australia with specific economic and industrial purposes that we know we will need using Australian datasets – some of the richest in the world. In doing so, it is vital that creative works are protected.
Provide centralised, funded research coordination around economic priorities, including an Australian LLM subsidised for priority industries and companies and supporting related model development and applications.
What Australia lacks at the moment is a centralised capacity to bring researchers together on a project such as this in a way that would be appropriately funded and organised around economic priorities: including an Australian LLM that would both be made available to priority industries and companies for subsidised use, and would also act as a central point to build knowledge and experience in model development that would have a flow on effect to other projects (including the building of smaller models; the adaptation of open-weight models; the development of tools and applications built on top of LLMs etc) structured around economic priorities. That is what our proposed coordination vehicle would provide.
Govern, store and process Australian data, especially sensitive government-held citizen and resident data, under Australian jurisdiction with privacy, security and public-interest safeguards.
the absolute principle that Australian data (especially sensitive data on citizens and residents that the government collects and holds) must be governed, stored and processed under Australian jurisdiction, with appropriate safeguards to protect privacy, security and the public interest.
Respect First Nations data sovereignty and governance rights, giving Aboriginal and Torres Strait Islander peoples meaningful control over cultural and intellectual data use, management and sharing.
Aboriginal and Torres Strait Islander peoples' rights and expectations regarding First Nations data sovereignty and governance must be respected. Aboriginal and Torres Strait Islander peoples should have meaningful control over the use, management and sharing of their cultural and intellectual data.
Assess the risks to Australian data sovereignty of US frontier AI labs training models in Australia as part of any approval process.
The Australian Government has expressed its desire for US-based frontier AI labs to train their models in Australia. Proper consideration of the risks this poses for Australian data sovereignty should be assessed as part of any approval process.
Host and process AI systems trained on Australian citizens’ data for their benefit onshore in Australian data centres, with safeguards against offshore access and secondary use.
AI systems trained on the data of Australian citizens for the benefit of Australian citizens must be hosted and processed onshore in Australian data centres, with safeguards against offshore access and secondary use.
Provide greater oversight and transparency of datasets and algorithms used in government AI systems, and protect personal information from being used to train AI without informed consent.
There must be greater oversight and transparency of the datasets and algorithms used in government AI systems and personal information must be protected from being used to train AI without informed consent.
Have government plan for clear, measurable AI benefits across suburbs and regions.
This means ensuring clear and measurable benefit to all Australians, which is why it is so important that the Government has a plan to ensure the benefits of AI are spread across Australia’s suburbs and regions.
Strengthen APS oversight, enforcement and independent auditing; require early worker and union consultation with a genuine opportunity to influence the approach.
stronger central oversight and enforcement powers and independent auditing to ensure consistent standards across the APS; strong consultation obligations with an opportunity for workers and their union to engage at the earliest possible stage with a genuine opportunity to influence the approach taken;
Require public disclosure of AI use cases, and strengthened independent monitoring, reporting and auditing of high-risk APS systems.
enforceable transparency obligations including public disclosure of AI use cases; strengthened independent monitoring, reporting and auditing of high-risk systems;
Keep human decision-makers responsible and accountable for government decisions, and prioritise APS workforce AI capability.
action to ensure human decision-makers must remain responsible and accountable for government decisions; government prioritisation of building workforce capability for the use of AI in the APS;
Give individuals affected by AI-supported government decisions accessible, timely and enforceable rights to review and appeal.
support for individuals impacted by AI- supported decisions to have accessible, timely and enforceable rights to review and appeal decisions
Prioritise in-house government AI capability and invest in developing or adopting locally developed solutions wherever possible.
government prioritisation of in-house capability and invest in the development or adoption of locally-developed AI solutions wherever possible.
Resource and empower the AI Safety Institute to actively protect Australians from harm, not only to monitor risks.
Unions continue to advocate for the AISI to be given appropriate resources and powers to actively protect Australians from harm – not just to monitor risks.
Give the Government power to license the deployment and development of AI used in Australia, as banks and superannuation funds are licensed.
The Government should have the power to license the deployment and development of AI used in Australia: similar to how we license banks and superannuation funds.
Empower the AI Safety Institute, alongside other regulators, to issue, review and potentially revoke licences where developers or deployers demonstrably breach law or regulations or fail effective systemic risk mitigation.
We are proposing to enhance the AI Safety Institute’s powers to create the capability to issue operating licences to AI developers and deployers of models and systems. The AI Safety Institute already works alongside other regulators, and it is expected that it would continue to do so in the issuing, review, and potential revoking of licences where a developer or deployer can be demonstrated to have breached law and regulations or not taken effective proactive measures to mitigate these risks on a systemic level.
Require early engagement with affected workers, communities and stakeholders, information and resources for meaningful participation, and evidence that their views shaped data-centre decisions.
Genuine consultation is more than informing people about decisions that have already been made. It requires engaging affected workers, communities and stakeholders early enough to influence the outcome, providing them with the information and resources necessary to participate meaningfully, and demonstrating how their views have shaped the final decision.
Use FMIA Community Benefit Principles as a model for AI-standards consultation, and protect regional communities when addressing urban concerns.
Unions have advanced these arguments in the context of the FMIA Community Benefit Principles and suggest these can function as a model for best practice community consultation for the AI Standards. We also particularly note that while the concerns of urban communities and quality of life are valid and should be taken into consideration, this should not be at the expense of regional communities.
Legislate AI standards for large data centres with a legislated enforcement mechanism including penalties for noncompliance.
Unions support this move to legislate rules around DC development, and call for a legislated enforcement mechanism being put in place for these standards including penalties for noncompliance.
Require all data centre projects to produce or procure new renewable energy generation equivalent to a minimum 110% of their facility's forecast energy demand.
Require that all DC projects should produce or procure new renewable energy generation equivalent to a minimum 110% of their facility’s forecast energy demand.
Retain the requirement that data centre projects create fair, safe, secure and well-paid jobs, and extend it through supply chains and subcontracting with an emphasis on long-term jobs.
Maintain the DC expectations requirement that DC projects create “fair, safe, secure and well-paid jobs for Australian workers”. It is absolutely vital that this is retained in the standards, is extended to apply throughout supply chains and subcontracting arrangements with an emphasis on the creation of long-term jobs.
Ensure data centres are built in locations where they are not competing with land otherwise suitable for housing.
Ensure that DCs are built in locations where they are not competing with land otherwise suitable for housing.
Mandate an apprentice ratio of 20% of the trades workforce for the development phase of all data centres and their connected renewable energy projects.
A mandated apprentice ratio of 20% of the trades workforce should apply to the development phase of all DCs and to their connected renewable energy projects.
Require data centre projects that fail to meet the apprentice ratios to pay into a training fund financing new or expanded TAFE and industry-led RTO training capacity.
Projects that fail to meet the required ratios should pay into a training fund that finances new or expands existing TAFE and industry-led RTO training capacity.
Require data centre developers and operators to pay their fair share of tax.
Include requirements that all DC developers and operators pay their fair share of tax.
Require a clear commitment that data-centre developments will not negatively affect communities’ quality of life.
Considering the scale and extent of the quality of life threats to working people evidenced internationally and increasingly present in Australia, working people need a clear and unequivocal commitment to ensuring that no DC development will negatively impact on their quality of life.
Require meaningful community consultation, with particular engagement of Aboriginal and Torres Strait Islander communities, as a precondition of approval for data centre development, by amending the data centre expectation.
As a result, the DC expectation should be amended to explicitly reference meaningful consultation to address community concern as a precondition of approval for DC development. There should be particular reference to the need to engage meaningfully and sensitively with Aboriginal and Torres Strait Islander communities.
Do not divert resources allocated to FMIA green manufacturing and other projects to accommodate data centres.
In the transition to AI standards, it is important that resources already allocated to FMIA-connected projects in green manufacturing and elsewhere are not diverted from these purposes to accommodate DC projects.
Site or incentivise data-centre loads where renewable or network capacity exists; require full transmission and network-cost contributions, demand flexibility and peak-load cooperation, and facility-level energy and emissions reporting.
This means securing new and additional renewable energy supply that exceeds forecast facility demand by at least 10% (to create a buffer for higher than anticipated demand), aligning or incentivising data centre loads to be sited in locations with existing renewable or network capacity, covering in full their share of transmission infrastructure and flow-on network costs, cooperating with energy regulators and suppliers to enhance demand flexibility and opportunities for peak load management, and sharing data on energy consumption, energy sources, and annual emissions. Data centre reporting must be made available at the facility level to ensure local benefits and impacts can be adequately assessed.
Require new renewable capacity for a data centre to reach Final Investment Decision before the data centre begins operation.
Either way, the new renewable capacity must reach Final Investment Decision (FID) before the DC begins operation.
Apply the new and additional renewable energy requirement to both on-grid and off-grid data centres.
The requirement around new and additional renewable energy should apply to both on-grid and off-grid DCs, to prevent developers from simply bypassing regulations by building a large new fleet of bespoke gas-fired power plants.
Allow renewable-origin offsets only during development of new renewables, with the shortest feasible transition to physical 110% supply and firm safeguards against perpetual reliance on cheap offsets.
While on-site renewables or a Power Purchase Agreement (PPA)-backed new renewable project is in development, DCs may need some access to Renewable Electricity Guarantee of Origin offsets, but they must not be allowed to rely on these offsets in perpetuity, which would undermine the entire point of the regulation: to rapidly bring new physical renewable infrastructure online. The transitional period allowable for developers to achieve the 110% benchmark with new and additional renewables (as opposed to REGOs or equivalent) must therefore be as short as feasible, with firm guardrails to ensure the intent of the scheme is not watered down by cheap offsets.
Require data centre companies to prove additionality of renewable energy, particularly for power purchase agreements.
Finally, DC companies must be required to prove additionality, particularly in the case of PPAs, to pre-empt the risk that developers take credit for renewables that were already going to be built regardless of the status of their DC project.
Consider incentives and preference for data centres in regions with high renewable curtailment to use existing capacity and reduce new transmission needs.
These requirements could also be paired with incentives for DCs built in regions with high rates of curtailment of renewable generation.
Require data centre developers to prioritise non-potable water and publicly disclose total projected water use, at least aggregated by water catchment.
Non-potable water use should be actively prioritised – not simply encouraged “where possible”. DC developers must also be required to publicly disclose total projected water use, at least aggregated by water catchment, if not at the level of individual facilities, to inform government water use and resiliency planning.
Require water-minimisation plans from the outset and continual improvements in data-centre water efficiency.
Australia is a drought-prone country and DC developers should be required to draft plans to minimise water use from the outset and engaging in continual improvement of their facilities to make water use more efficient.
Consider updating the data-centre research, innovation and local-capability expectation to support a national pro-worker sovereign AI programme.
We encourage the Government to consider updating this expectation to support a national programme to build Pro-Worker Sovereign AI capability in Australia.
Support MEAA’s Copyright Plus framework combining remuneration for uncompensated workers, a cultural fund for public entertainment and cultural experiences, and transparency and accountability mechanisms.
MEAA is advocating for a ‘Copyright Plus’ framework, combining an equitable remuneration payment model for workers whose content has been used without compensation with a cultural fund that supports the development and delivery of public entertainment and cultural experiences, alongside a range of other mechanisms designed to facilitate greater transparency and accountability across the sector. Australian Unions refers the Committee to the MEAA submission to this Inquiry, which we strongly support.
Extend the Digital Duty of Care throughout the technology stack, including generative AI chatbots and image generators.
As a result, Australian Unions strongly support the Government’s proposed Digital Duty of Care and its inclusion of services throughout the tech stack, including but not limited to ‘generative AI services (including chatbots and image generators)’.
Establish explicit expectations, standards and compliance requirements with strong regulatory oversight and enforcement powers, under a Digital Duty of Care placing responsibility for preventing harm with those who design, deploy and profit from digital systems.
Unions believe Government should establish explicit expectations, standards and compliance requirements, supported by strong regulatory oversight and enforcement powers. A Digital Duty of Care should ensure that responsibility for preventing harm sits with those who design, deploy and profit from digital systems.
Pursue complementary state-level reforms to criminalise the production and distribution of explicit deepfake material.
Unions also note the need for complementary reforms at the state level to criminalise the production and distribution of explicit deepfake material, such as those currently being proposed in NSW.
In the joint supplement, establish a government-coordinated Sovereign AI Coordination Vehicle bringing together unions, researchers, industry and superannuation to build Australian AI infrastructure and models, including foundation models, with worker consultation and secure jobs.
we are jointly calling for a Sovereign AI Coordination Vehicle to be established by the Australian Government. This will represent a government-coordinated effort bringing together unions, universities and industry experts, the superannuation sector, and the Australian tech industry to enhance our control over AI infrastructure and build our own AI models, including Australian foundation models, deployed with worker consultation and suited to the needs of our economy creating secure, safe, high- quality jobs for the future.
In the joint supplement, establish a standing corporate-union-government body with university/public-research board representation, undertaking fundamental research and commercialisation support for national priorities, global competitiveness and safe, secure, well-paid jobs.
A. Will be a standing tripartite body established with foundational corporate members, union representatives, and the Australian Government, with the research sector represented on the board through universities and publicly funded research organisations, to undertake the fundamental research and commercialisation support required to create global competitive advantage for the Australian AI industry while supporting national economic priorities and creating safe, secure, and well-paid jobs. This is critical if our companies are to develop and pursue global market opportunities.
In the joint supplement, establish the vehicle as a not-for-profit with a tripartite board; retain publicly funded or public-data intellectual property in Australia for public benefit, potentially housing an AI R&D corporation.
B. Will be established as a not-for-profit company with a tripartite board, so that it can move with the speed and commercial flexibility required to deliver a competitive advantage for Australian industry, while ensuring that intellectual property generated with public funding or public data remains in Australia and returns value to the Australian public. It could take the form of, or house, an AI Research and Development Corporation modelled on the rural research and development corporations.
In the joint supplement, invest in Australian-controlled compute, data, secure development/testing, engineering and AI models; commission compute through national research infrastructure and Australian-owned neocloud providers, allocated by national merit as one pool.
C. Will invest in Australian-controlled compute, data infrastructure, secure development and testing environments, engineering support, and a portfolio of Australian AI models and applications. Compute would be commissioned rather than built from scratch, operated by Australia’s existing national research infrastructure providers and Australian-owned neocloud providers, and allocated through the established national merit process, so that capacity forms a single national pool rather than fragmenting across institutions, states and vendors.
In the joint supplement, admit new corporate partners, including investment funds, only with ongoing funding, mission/value alignment and acceptance of Australian control over compute, data and intellectual property.
New corporate partners can join, including investment funds, based on ongoing financial commitment, alignment with values and mission, and acceptance that the Vehicle’s compute, data and intellectual property remain under Australian control.
In the joint supplement, fund and coordinate vendor-neutral, transferable training across the AI capability chain, including builders, infrastructure staff, technicians, data, security, assurance and domain specialists, tradespeople and AI-using workers.
D. Will invest in and assist in coordinating training across the full AI capability chain: AI researchers and builders, research software and infrastructure engineers, technicians, data stewards, security specialists, AI assurance evaluators, domain experts, tradespeople and workers who will use AI. It will bring world-class AI skills to the Australian people. Training will be vendor-neutral, building transferable capability rather than familiarity with particular products.
In the joint supplement, fund the coordination vehicle through Commonwealth investment alongside contributions from the AI infrastructure sector.
E. Will be funded through Commonwealth investment alongside contributions from the AI infrastructure sector.
Require developers seeking Commonwealth/state facilitation to reserve defined shares of installed AI compute for the national pool, in kind or by trade, and fund people/software/data; bind compute controllers through operator-to-tenant contracts.
Data centre and AI infrastructure developers seeking Commonwealth or state facilitation should be required to reserve a defined share of installed AI compute for the national pool, discharged in kind or by trade to another provider, and to make an AI Capability Contribution that funds the people, software and data infrastructure needed to use it. These obligations should bind the party that controls the compute, with contractual flow-down from data centre operators to their tenants.
In the joint supplement, exclude free tokens, expiring credits and vendor-specific training from obligations, and require major AI companies doing substantial Australian business to maintain genuine onshore R&D.
Free tokens, expiring credits and vendor-specific training should not count toward them. Major AI companies doing substantial business in Australia should also maintain a genuine research and development presence onshore.
In the joint supplement, coordinate mission-driven Australian AI models and applications in areas of need and comparative advantage, with First Nations data governed by Indigenous Data Sovereignty and CARE principles.
The coordination vehicle would oversee the development of Australian AI models, from domain-specific to foundation models, and seed the development of a portfolio of mission- driven projects in areas of Australian need and comparative advantage, including health, energy, climate, critical minerals, agriculture, manufacturing, defence, public services, education and Australian language and culture, with any use of First Nations data governed by Indigenous Data Sovereignty and the CARE principles.
In the joint supplement, bring companies, workers, research organisations, universities, TAFEs, investors and government together around defined project outcomes, including secure, safe and well-paid jobs.
Projects should bring together companies, workers, research organisations, universities, TAFEs, investors and government around defined outcomes, including supporting and creating secure, safe, and well-paid jobs.
In the joint supplement, preserve public control where public money and data are central, use openness where safe and appropriate, and restrict deployment where sensitivity requires.
Australia should adopt the same principle: public control where public data and public money are central, openness where safe and appropriate, and restricted deployment where sensitivity requires it.
In the joint supplement, use sovereign compute and a single national pool to provide startups GPU access.
The Coordination Vehicle would invest in vital AI infrastructure, such as the development of sovereign AI compute to support the development of locally customised AI models and provide GPU access for startups through the single national pool described above.
In the joint supplement, consider regional AI factories connecting data, compute and talent with universities, businesses and investors to accelerate AI businesses.
The Coordination Vehicle could establish regional AI factories: AI hubs that accelerate new AI businesses by connecting data, compute capacity, and talent with universities, businesses, and investors.
Have government technology procurement explicitly seek to keep value created in Australia within Australia and prioritise well-paid, secure and safe jobs throughout the supply chain.
Building on the Buy Australia Plan, government tech procurement should explicitly seek to keep value created in Australia within Australia, and to prioritise the creation of well-paid, secure, and safe jobs throughout the supply chain.