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 97

CAM Initiative

Submitter group: Research institutes.

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.

Proposes a prosperity framework connecting capability discovery, funding, building, scaling, retention and reinvestment, building on existing programs. Recommends direct-access capability support, a protected civilian Sovereign Investment Fund, extreme-wealth thresholds, domestic robotics and distribution, and strategic-asset retention. Proposes a copyright remuneration registry that preserves licensing and does not itself permit use of protected material.

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 mandatory cash co-contribution at pre-seed and capability-discovery stages defeats public funding’s purpose of overcoming financing constraints.

    At the pre-seed and capability-discovery stage, mandatory cash co-contribution in a funding application constitutes a policy-design failure. It prohibits access to public capital intended to overcome financing constraints on an applicant who does not already possess capital.

  2. The submission recommends that the Commonwealth establish a long-term public investment fund to retain and reinvest part of the wealth created through AI and automation.

    The Commonwealth should establish a long-term public investment fund that allows Australia to retain and reinvest part of the wealth created through AI, automation and other strategically important productive assets.

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

  3. The submission argues that the problem is ensuring Australian firms can reach global scale without ownership, decision rights and returns progressively moving offshore.

    Australia's problem is ensuring Australian firms can reach global scale without the ownership, decision rights and compounding economic returns progressively moving offshore.

  4. The submission argues that at the scale stage the question is whether existing investment mechanisms are deployed at sufficient scale with strategic-retention objectives, not whether such mechanisms exist.

    At Scale, the policy problem is no longer the absence of investment mechanisms or how to fund a prototype. It is whether existing Australian mechanisms can be deployed at sufficient scale, with appropriate strategic-retention objectives, to finance and retain firms and infrastructure large enough to become enduring capability

  5. Argues that regulation can improve safety, choice and accountability without itself creating domestic capacity to build, own or export discovery and monetisation systems.

    Regulation can improve safety, choice and accountability; it does not by itself create domestic capacity to build, own or export the systems through which discovery and monetisation occur.

  6. The submission proposes establishing extreme wealth thresholds, stating the purpose is not to constrain ordinary wealth, successful businesses, home ownership or retirement savings.

    Australia should also establish extreme wealth thresholds. The purpose is not to constrain ordinary wealth, successful businesses, home ownership, retirement savings or reasonable intergenerational inheritance.

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 an Australian AI Capability Stream before institutional backing, incorporation or outside capital; expand fellowship pathways for demonstrated capability; require no matched funding at feasibility/capability-proof stage and incorporation only after selection where contracting requires.

    The Commonwealth should establish a standing Australian AI Capability Stream that can identify and develop emerging capability before people or teams already have institutional backing, a company or outside capital. It should expand existing fellowship-style pathways for applicants who can demonstrate capability through projects, open-source work, hackathons or other practical evidence. Entry at feasibility or capability-proof stage should not require matched funding, and incorporation should be required only after selection where needed for contracting.

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

  2. Provide capability-stream support including paid research/build time, compute/API and prototype/facility access, mentorship, peer review, legal/research-governance support and remote participation.

    Support may include paid research or build time, compute or API access, prototype or technical-facility access, mentorship, peer review, legal and research-governance support and remote participation.

  3. Consider CSIRO, working with the Australian AI Safety Institute and other relevant institutions, as a possible host or delivery partner for the Capability Stream.

    CSIRO, working with the Australian AI Safety Institute and other relevant institutions, should be considered as a possible host or delivery partner.

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

  4. Select capability-stream applicants by demonstrated work, Australian connection, merit and a credible work program, aiming to convert capability into Australian research, employment, firms, IP and services.

    Selection should turn on demonstrated work, Australian connection, merit and a credible work program, with the aim of converting emerging capability into Australian research, employment, firms, intellectual property and services. [3-5]

  5. Connect existing public investment into an AI strategic-capital pathway, using the National Reconstruction Fund Corporation and existing mechanisms first for Australian AI, digital and automation capability moving from commercialisation toward scale and strategic importance.

    At the Fund stage, Australia should connect existing public-investment machinery into a clearer AI strategic-capital pathway. The National Reconstruction Fund Corporation and other existing mechanisms should be used first to invest in promising Australian AI, digital and automation capability as it moves from commercialisation toward scale and strategic i mportance.

  6. Where existing mandates, allocations or available capital are insufficient, extend or supplement the existing investment machinery rather than duplicate it.

    Where existing mandates, allocations or available capital are insufficient, government should extend or supplement that machinery rather than duplicate it.

  7. Establish a long-term Sovereign Investment Fund retaining AI, automation and strategic-asset wealth; permit wealth-overflow receipts and specifically authorised public-value streams, including designated government equity/infrastructure returns, royalties and licensing income.

    The Commonwealth should establish a long-term public investment fund that allows Australia to retain and reinvest part of the wealth created through AI, automation and other strategically important productive assets. The Fund could receive wealth-overflow receipts and other specifically authorised public-value streams, including designated returns from government equity, infrastructure investments, royalties or licensing arrangements.

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

  8. Use Fund returns over time for productive capability, communities affected by technological transition and future generations’ share of automation benefits.

    Its purpose should be to convert temporary or concentrated economic gains into permanent public wealth. Investment returns should be used over time to strengthen Australian productive capability, support people and communities affected by technological transition, and ensure future generations share in the economic benefits created by automation.

  9. Protect the Fund by statute from general revenue or unrelated uses, including military expenditure; keep it independent of day-to-day government spending.

    The Fund should operate independently of day-to-day government spending, with clear statutory protections preventing its assets from being absorbed into general revenue or diverted to unrelated purposes, including military expenditure.

  10. Where public support materially contributes to a valuable asset, the Commonwealth should be capable of retaining part of the resulting commercial value through the relevant instrument.

    Where public investment, infrastructure, procurement, research, concessional finance or strategic support materially contributes to the creation of a valuable asset, the Commonwealth should be capable of retaining part of the resulting commercial value through the relevant instrument.

  11. Keep existing investment returns within statutory mandates; transfer a share of public productive returns to the Sovereign Investment Fund only with express parliamentary authorisation.

    Existing investment vehicles should retain and reinvest returns according to their statutory mandates. Where Parliament determines that a share of public productive return should also support the proposed Sovereign Investment Fund, that transfer should be expressly authorised rather than assumed.

  12. Establish extreme wealth thresholds, not intended to constrain ordinary wealth, successful businesses, home ownership, retirement savings or reasonable inheritance.

    Australia should also establish extreme wealth thresholds. The purpose is not to constrain ordinary wealth, successful businesses, home ownership, retirement savings or reasonable intergenerational inheritance.

  13. Above an extreme-wealth threshold, progressively return a proportion of excess wealth to the common productive system; follow beneficial ownership, preserve accrued obligations after offshore transfer or departure, and avoid forced business liquidation.

    Above the threshold, a proportion of attributable excess wealth should progressively overflow back into the common productive system. The mechanism should follow beneficial economic ownership, survive offshore transfer or departure once an obligation has accrued, and avoid forced liquidation of productive enterprises.

  14. Calibrate to national wealth distribution rather than fixing an arbitrary dollar amount; illustratively test five-standard-deviation thresholds above the Australian log-wealth median/centre, with a billionaire-scale activation floor.

    A wealth threshold should be calibrated against national wealth distribution rather than fixed permanently at an arbitrary dollar figure. As an illustrative example, a reference model would test extreme statistical thresholds such as “ 5 standard deviations above an Australian median/centre of the log-transformed wealth distribution ” with a billionaire-scale activation floor.

  15. Require Treasury modelling and periodic recalibration of statistical wealth thresholds and the activation floor.

    The statistical threshold and activation floor would require Treasury modelling and periodic recalibration against the Australian wealth distribution.

  16. Prevent ownership-preserving transfers from resetting wealth calculations; preserve accrued obligations after departure; trigger valuation, preservation or settlement on departure, offshore transfer or restructuring, with credits against genuine double taxation.

    Moving shares into another company, trust, foundation or foreign account should not reset the calculation where the same person continues to own or control the underlying economic interest. Similarly, a person who leaves Australia after an overflow obligation has accrued should not be able to extinguish that obligation simply by changing residence. Departure, offshore transfer and major restructuring should therefore trigger valuation, preservation or settlement rules, while appropriate credits should prevent genuine double taxation.

  17. Avoid unnecessary sale of productive businesses; for illiquid wealth, allow deferral, security over ownership, partial economic interests or progressive payment on dividends, sales, buybacks, refinancing, inheritance or other liquidity events.

    The mechanism should also avoid forcing productive businesses into unnecessary sale. Where wealth is concentrated in an operating company or other illiquid asset, payment could be deferred, secured against the ownership interest, satisfied partly through an economic interest, or progressively met when dividends, sales, buybacks, refinancing, inheritance or other liquidity events occur.

  18. Protect the Fund’s principal and govern distributions through transparent rules, rather than annual political discretion; prevent general-revenue substitution or diversion into military capability.

    The principal should be protected, distributions should be governed by transparent rules rather than annual political discretion, and the founding legislation should prevent the Fund from becoming a substitute source of general revenue or being diverted into military capability.

  19. Study an Australian-owned or meaningfully Australian-equity AI-native discovery/creator-distribution platform; protect editorial independence, privacy, competition, expression and independence from ministerial control, with commercially disciplined public participation.

    The Commonwealth should commission a commercial and governance feasibility study for an Australian-owned or meaningfully Australian-equity discovery and creator-distribution platform for an AI-native media environment. The platform should be examined not only as commercial infrastructure, but as part of Australia’s capability-discovery system: giving emerging Australian researchers, creators, projects and businesses a stronger pathway to visibility, audiences, collaboration and market access than reliance on foreign recommendation systems alone. It should also seek to retain more of the distribution, advertising and creator value generated by Australian activity within the Australian economy. Any public participation should be commercially disciplined while protecting editorial independence, privacy, competition, freedom of expression and independence from ministerial control.

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

  20. Examine the proposed platform as capability-discovery infrastructure for researchers, creators, projects and businesses, improving visibility, audiences, collaboration and market access while retaining Australian distribution, advertising and creator value.

    The platform should be examined not only as commercial infrastructure, but as part of Australia’s capability-discovery system: giving emerging Australian researchers, creators, projects and businesses a stronger pathway to visibility, audiences, collaboration and market access than reliance on foreign recommendation systems alone. It should also seek to retain more of the distribution, advertising and creator value generated by Australian activity within the Australian economy.

  21. Structure public platform participation as investment rather than perpetual subsidy, retaining public commercial benefit if successful; protect expression, editorial independence, privacy, data protection and competition, and prevent ministerial control over lawful content.

    Public participation should be structured as investment rather than perpetual subsidy. If a sovereign-backed Australian platform succeeds, the public should retain part of the commercial benefit. Governance should protect freedom of expression, editorial i ndependence, privacy, data protection and competition, and should explicitly prevent ministerial control over lawful content.

  22. Through the Attorney-General’s Department and CAIRG, preserve licensing, establish a Copyright & AI Claims Registry, and require commercial AI suppliers/deployers above de minimis Australian-market thresholds to register and contribute proportionately to creator remuneration.

    The Commonwealth should develop, through the Attorney-General’s Department and CAIRG, a commercial AI copyright-reconciliation framework that preserves direct and voluntary collective licensing, establishes a government-backed Copyright & AI Claims Registry, and requires covered commercial AI deployers or suppliers above de minimis Australian market thresholds to register and make proportionate contributions to a creator-remuneration pool.

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

  23. Give verified direct-licensing payments transparent credits and allocate verified claims through an indexed Contribution Value Schedule; registration or contribution must not itself grant permission to use protected material.

    Verified direct-licensing payments should receive transparent credits, and verified claims should be allocated through an indexed Contribution Value Schedule. Registration or contribution must not itself grant permission to use protected material.

  24. Have known upstream rights and provenance uncertainty follow material fine-tunes, distillations and successor models, with safe harbours for genuinely small and non-commercial open-source projects.

    Known upstream rights and provenance uncertainty should follow material fine-tunes, distillations and successor models, with safe harbours for genuinely small and non-commercial open-source projects.

  25. Retain existing direct/collective and statutory licences, lawful exceptions, open licences and public-domain use; limit the new mechanism to opaque commercial use without authorising ingestion or extinguishing copyright obligations.

    Direct and collective licences, statutory licences, lawful exceptions, open licences and public-domain material should continue within their existing legal scope. A new claims mechanism should address the residual problem of opaque commercial use; participation in it should not itself authorise i ngestion or extinguish ordinary copyright obligations.

  26. Have the Attorney-General’s Department administer the Registry, with Australian AI Safety Institute technical support on provenance, model lineage and provider-reporting standards.

    The Attorney-General’s Department should administer a dedicated Copyright & AI Claims Registry, with technical support from the Australian AI Safety Institute on provenance, model-lineage and provider-reporting standards.

  27. Make rights-holder registration free or low-burden, recording ownership, source/publication evidence, licensing status and available evidence of likely AI use.

    Rights-holder registration should be free or l ow-burden and record ownership, source or publication evidence, licensing status and any available evidence of likely AI use.

  28. Require covered providers to keep regulator-accessible corpus, acquisition, licensing and synthetic/distilled-lineage records, with proportionate trade-secret protections.

    Covered providers should maintain regulator-accessible records of material corpus classes, acquisition pathways, licensing status and known synthetic or distilled lineage, with proportionate protections for legitimate trade secrets.

  29. Separate collection from allocation; base above-threshold commercial-provider contributions on Australian AI-attributable gross revenue, using active users, enterprise reach, training/inference capacity and model capability as scale and anti-avoidance modifiers.

    The scheme should separate collection from allocation. On the provider side, commercial AI suppliers or deployers participating in the Australian market should register above defined de minimis thresholds and contribute according to Australian AI-attributable gross revenue, with active users, enterprise reach, training or inference capacity and model capability used as scale and anti-avoidance modifiers.

  30. Consider contribution tiers: none below threshold, reduced/transitional for emerging commercial providers, standard revenue-linked for established providers, progressively higher for systemic/frontier providers; permit reduction/exemption for non-commercial research/public-interest use without material commercial revenue.

    Contribution obligations could be organised into provider tiers: entities below the threshold would have no additional contribution obligation; emerging commercial providers would pay a reduced or transitional rate; established commercial providers would pay a standard revenue-linked contribution; and systemically significant or frontier providers would face a progressively higher contribution reflecting their scale, capability and dependence on large training corpora. Non-commercial research and public-interest uses that do not generate material commercial revenue should be eligible for reduced contributions or exemption.

  31. Distinguish output resemblance from causal-use evidence; similarity alone should not establish ingestion or model lineage.

    Output similarity should not, by itself, establish ingestion or model lineage. Generative systems can independently produce similar language, structures or solutions in response to similar prompts, common source material or constrained problem spaces. The claims process should therefore distinguish resemblance from evidence of causal use.

  32. Verify rights, provenance and contribution evidence without requiring creators to prove hidden ingestion; shift the exclusion/non-use/licensing burden to systemic/frontier providers where rights, credible provenance and plausible training-source connection are established.

    The Registry should verify ownership, provenance and contribution evidence, but creators should not be required to prove hidden ingestion within frontier models that obscure or de-identify source material. Where a claimant can establish rights, credible provenance and a plausible connection to a provider’s training sources, systemically significant or frontier providers should bear the burden of showing that the material was excluded, not used or already covered by an applicable licence.

  33. Limit the presumption to remuneration eligibility, without establishing infringement or permission; allocate verified claims through a published, indexed Contribution Value Schedule.

    This presumption would apply only to eligibility for remuneration. It would not establish copyright infringement or grant permission to use protected material. Verified claims should then be allocated through a published, indexed Contribution Value Schedule. Table 1 sets out an indicative structure.

  34. In the indicative remuneration schedule, assess rights, permission, licensing restrictions and prior direct licensing to establish eligibility and prevent double payment, crediting verified licensing payments.

    Rights and permission Whether the claimant holds relevant Establishes eligibility and rights; whether use was licensed, prevents double payment. restricted or unlicensed; and whether Verified direct-licensing direct licensing has already occurred. payments receive a credit.

  35. In the indicative schedule, assess traceable independent human contribution and provenance; meaningful demonstrated human contribution should retain value despite AI assistance.

    Human origin and Evidence that the material contains Strong provenance increases provenance i ndependent human intellectual or confidence in the claim. AI assistance does not Component Assessment Effect on remuneration creative contribution and can be automatically reduce value reliably traced to its source. where meaningful human contribution remains demonstrable.

  36. Indicatively, U0 inputs without demonstrated independent contribution or wholly redundant synthetic material receive no utility premium beyond surviving rights; general-purpose U1 inputs in large/saturated source classes receive low weighting.

    Utility Band: U0 — No No demonstrated independent No utility premium beyond any demonstrated utility contribution to the relevant AI system, surviving rights entitlement. or wholly redundant synthetic material. Utility Band: U1 — General-purpose material within a Low utility weighting. General contribution l arge or saturated source class.

  37. Indicatively, U2 specialist/domain-specific inputs providing identifiable additional value receive moderate weighting; scarce, structured, high-quality, curated or independently validated U3 inputs receive high weighting.

    Utility Band: U2 — Distinctive, specialist or Moderate utility weighting. Differentiated domain-specific material that provides contribution i dentifiable additional value. Utility Band: U3 — Scarce, structured, high-quality, High utility weighting. High-value curated or independently validated contribution material.

  38. Give critical safety, evaluation, specialist or living-dataset U4 contributions the highest utility weighting, potentially supplemented by stewardship/dependency weighting, where model performance materially depends on them.

    Utility Band: U4 — Material important to safety, Highest utility weighting, Critical contribution evaluation, specialist capability, living potentially supplemented by datasets or other functions on which stewardship or dependency model performance materially weighting. depends.

  39. In the indicative schedule, add stewardship/currency weighting for continuing maintenance, updating, validation, correction or curation that preserves or increases value.

    Stewardship and Ongoing maintenance, updating, Additional weighting where currency validation, correction or curation of continuing human effort material. preserves or increases the material’s value.

  40. In the indicative schedule, add dependency/scarcity weighting for scarce or materially relied-upon inputs, assessing substitutability and material system dependence.

    Dependency and Whether equivalent material is readily Additional weighting for scarce scarcity substitutable or whether a system or materially relied-upon materially depends on a scarce source, i nputs. dataset or body of expertise.

  41. Set remuneration values in a transparent, reviewable indexed schedule outside primary legislation; periodically review rates as practices, markets and evidence change, publishing modelling of distributions across rights-holder classes.

    The monetary value attached to these units should be set through a transparent, reviewable and indexed schedule rather than hard-coded into primary legislation. Rates should be periodically reviewed as AI training practices, markets and evidence capabilities change, with published modelling showing expected distributions across different classes of rights holders.

  42. Use the National Reconstruction Fund, procurement, R&D and commercialisation programs, demonstration projects and targeted manufacturing support to strengthen Australian robotics, automation, component, software, integration, maintenance and control-system capability.

    The Commonwealth should use the National Reconstruction Fund, procurement, R&D and commercialisation programs, demonstration projects and targeted manufacturing support to strengthen Australian robotics, automation, component, software, integration, maintenance and control-system capability.

  43. Where technology is imported, separate the surrounding value chain so Australians can own productive assets and provide finance, software, infrastructure, maintenance, servicing, distribution and other downstream functions.

    Where the underlying technology is imported, policy should deliberately separate the surrounding value chain so that Australian households, businesses and institutions can still own productive assets and Australian firms can provide financing, software, infrastructure, maintenance, servicing, distribution and other downstream functions.

  44. Use market design, procurement and ownership safeguards where necessary to prevent automated industries becoming vertically integrated offshore ecosystems in which most productive income flows overseas.

    Market design, procurement and ownership safeguards should be used where necessary to prevent automated industries from becoming vertically integrated offshore ecosystems in which Australian users bear the transition costs while most productive income flows overseas.

  45. Reward genuine Australian value-add, capability and ownership in public support, rather than final assembly, rebadging or passive resale of foreign technology.

    Public support should reward genuine Australian value-add, capability and ownership rather than final assembly, rebadging or passive resale of foreign technology.

  46. Where public support is provided, require genuine Australian value-add: local IP, components/subsystems, accredited training/apprenticeships, maintenance/repair capacity, export potential and durable supply-chain participation, rather than final assembly or rebadging alone.

    Where public support is provided, stronger terms should attach to genuine Australian value-add: locally created intellectual property, component or subsystem capability, accredited training and apprenticeships, maintenance and repair capacity, export potential and durable supply-chain participation rather than final assembly or rebadging alone.

  47. Regulate autonomous passenger fleets for broad participation; consider beneficial-owner vehicle limits, related-company aggregation rules against avoidance, and protected access for individuals, small businesses and independent Australian fleet owners.

    Commercial autonomous fleets should operate within a regulated Australian passenger-transport system, with ownership and fleet-concentration rules designed to preserve broad participation. This could include limits on the number or proportion of vehicles controlled by a single beneficial owner within a market, aggregation rules preventing those limits from being avoided through related companies, and protected access for individuals, small businesses and other independent Australian fleet owners.

  48. Consider a common Australian autonomous-mobility platform for bookings, payments, safety, accessibility and service standards, potentially regulated or publicly owned so different vehicle owners compete without manufacturer or dominant-fleet control.

    Australia should also consider establishing or supporting a common Australian autonomous-mobility platform that performs the distribution function currently provided by services such as Uber: connecting passengers with available vehicles, managing bookings and payments, and applying common safety, accessibility and service requirements. The platform could operate as regulated or publicly owned infrastructure through which many different vehicle owners compete, rather than allowing the vehicle manufacturer or a dominant fleet owner to control the customer network as well.

  49. Pair automation-ownership measures with transition measures such as wage-loss protection, portable learning, apprenticeships, retraining, regional job creation, public or community employment and stronger income support.

    The objective should be to create distributed opportunities for Australian households, workers and businesses to own, operate and earn income from productive automated assets, alongside transition measures such as wage-loss protection, portable learning, apprenticeships, retraining, regional job creation, public or community employment and stronger income support.

  50. Identify AI and digital infrastructure firms and assets of strategic national importance and establish a retention pathway as they scale.

    The Commonwealth should identify AI and digital infrastructure firms and assets of strategic national importance and establish a retention pathway as they scale.

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

  51. Where substantial growth capital is required, allow government to co-invest alongside Australian superannuation and private capital, or retain strategic equity or rights.

    Where substantial growth capital is required, government should be able to co-invest alongside Australian superannuation and private capital, or retain strategic equity or rights.

  52. Allow the Sovereign Investment Fund, where consistent with its eventual statutory mandate, to consider direct investment or co-investment in a strategically important Australian-origin asset offered for sale.

    Where consistent with its eventual statutory mandate, the Sovereign Investment Fund should also be able to consider direct investment or co-investment in a strategically important Australian-origin asset offered for sale.

  53. Keep foreign-investment conditions available where necessary.

    Foreign-investment conditions should remain available where necessary.

Read the original

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

Open in the interactive guide