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 209

Australasian Music Publishers' Association Limited (AMPAL)

Submitter group: Creative sector.

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.

Argues that Australia's existing copyright framework can accommodate AI and that copyright should not be weakened, while supporting responsible AI investment. Opposes a text and data mining exception, compulsory or extended collective licensing, opt-out requirements and AI-specific safe harbours; supports voluntary licensing, work-level transparency and enforcement.

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 existing copyright rules should apply to AI uses controlled by rightsholders, with uncertainty addressed through guidance, licensing practice and judicial decisions.

    Existing copyright rules should continue to apply when AI developers copy, store, reproduce or otherwise use copyright-protected material in ways controlled by the rightsholder. Uncertainty about their application to particular technologies or circumstances can be addressed through guidance, licensing practice and judicial decisions

  2. Describes Suno’s licensed-content v6 launch and other initiatives as evidence that negotiated licensing can operate at scale.

    On 9 September 2026, Suno released its new v6 models, developed with Warner Music Group, BMG and Believe using licensed content 9. Together with the other initiatives outlined above, the launch provides further evidence that negotiated licensing can operate at scale, make substantial repertoires available and support the development of advanced commercial AI models.

  3. Argues that a long tail of works does not in itself justify compulsory or extended collective licensing, presumed authority or delay in licensing identifiable rightsholders.

    A ‘long tail’ in itself does not justify compulsory or extended collective licensing, presumed authority or delay in licensing known and identifiable rightsholders.

  4. Argues that rightsholders should not be required to opt out, and that the absence of an opt-out signal should not be treated as consent.

    AMPAL opposes any requirement for rightsholders to opt out, register objections, use robots.txt or deploy other technological measures to prevent unauthorised use. Rightsholders may choose to use such measures, but their absence must not be treated as consent.

  5. Argues that transparency, consent, attribution and remuneration are complementary requirements, and that disclosure does not authorise use.

    Disclosure does not authorise use. Consent, transparency, attribution and remuneration are complementary requirements; none is a substitute for another.

  6. Argues that AI investment and sovereign capability should proceed alongside, not at the expense of, creative industries and copyright.

    AMPAL supports responsible AI investment and sovereign capability alongside, not at the expense of, Australia’s creative industries. Australia need not choose between investment and copyright.

Recommendations

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

  1. Confirm copyright applies to AI development and operation, prioritise enforcement, and require providers to demonstrate provenance and lawful sourcing of model-development material, including third-party datasets.

    confirm that Australian copyright applies to AI development and operation, prioritise enforcement, and require AI providers to demonstrate the provenance and lawful sourcing of material used in model development, including material obtained through third-party datasets.

  2. Reject any text and data mining exception, compulsory or extended collective licence, new AI-specific safe harbour, presumed mandate or requirement for rightsholders to opt out or deploy technology to preserve rights.

    reject the implementation of any form of text and data mining (TDM) exception, compulsory or extended collective licence, new AI-specific safe harbour, presumed mandate or requirement for rightsholders to opt out or deploy technology to preserve existing rights.

  3. Support direct and expressly mandated collective licensing on mutually agreed terms, without delaying identifiable-repertoire licensing while broader long-tail questions are considered.

    support direct and expressly mandated collective licensing on mutually agreed terms, without delaying licensing of identifiable repertoire while broader ‘long-tail’ questions are considered.

  4. Introduce proportionate, machine-readable, work-level transparency throughout model development, including third-party datasets and web-scraped material, and require preservation of copyright-management information.

    introduce proportionate, machine-readable, work-level transparency throughout model development, including for material obtained through third-party datasets or collected by web- scraping tools, and require the preservation of copyright-management information so rightsholders can identify unauthorised uses and enforce their rights.

  5. Require proponents of intervention to demonstrate the nature and extent of any market failure, the necessity of intervention and why less intrusive voluntary solutions would be insufficient.

    require proponents of intervention to demonstrate the nature and extent of any market failure, the necessity of intervention and why less intrusive voluntary solutions would be insufficient.

  6. Ensure reforms, incentives, agreements and procurement protect licensing markets, satisfy treaty and constitutional obligations and are subject to full public consultation.

    ensure reforms, incentives, agreements and procurement protect licensing markets, satisfy treaty and constitutional obligations and are subject to full public consultation.

  7. Ensure government intervention does not undermine emerging licensing markets, deal values or Australian rightsholders’ negotiating position in global arrangements.

    Government intervention must not undermine emerging licensing markets, deal values or Australian rightsholders’ negotiating position in global arrangements.

  8. Treat legislative intervention as premature until major AI developers demonstrate sustained attempts to license identifiable repertoire through available direct and expressly mandated collective channels; prioritise enforcement over new immunities.

    Legislative intervention is premature while major AI developers have not demonstrated sustained attempts to license identifiable repertoire through available direct and expressly mandated collective channels. The priority should be enforcement, not new immunities.

  9. Before considering legislation, require long-tail proponents to define the repertoire, separate protected and public-domain works, identify genuinely unlocatable rightsholders, and establish necessity and added model value beyond voluntarily licensable repertoire.

    Before considering legislation, proponents should define any claimed licensing ‘long tail’, distinguish copyright- protected repertoire from musical works already in the public domain, identify which rightsholders genuinely cannot be located and establish whether access to that repertoire is necessary for viable AI models, and the value that would be added to the models beyond licensable repertoire under direct or collective agreements.

  10. Do not treat absent opt-out signals as consent or let post-training opt-outs merely remove payment while the provider retains the benefit of an incorporated work.

    An opt-out exercised only after a work has been used to train a model may provide no meaningful control if the provider is unable to reverse or remove the effects of that use. If the work remains incorporated in the model, opting out must not simply result in the rightsholder losing payment while the provider retains the benefit of the use. Prior informed permission provides greater certainty for both rightsholders and AI developers. The absence of an opt-out signal is not permission.

  11. Base voluntary licensing models on appropriately scoped express mandates, preserve rightsholder choice and direct agreements, and provide transparent governance, reporting, valuation, distribution and safeguards.

    Any voluntary model should rest on appropriately scoped express mandates, preserve rightsholder choice and direct agreements, and provide transparent governance, reporting, valuation, distribution and safeguards.

  12. Support voluntary licensing by bringing rightsholders and AI providers together and encouraging compatible data and reporting standards.

    Government can support voluntary licensing by bringing rightsholders and AI providers together, and encouraging compatible data and reporting standards that make works and rightsholders easier to identify across different systems.

  13. Involve AMPAL directly in developing any copyright or Australian music-sector framework.

    AMPAL should therefore be directly involved in developing any copyright or specific Australian music-sector framework.

  14. Reject other limitations of liability for providers’ own copying or use of protected material in model development.

    AMPAL opposes any new AI-specific safe harbour or other limitation of liability for a provider’s own copying or use of protected material in model development.

  15. Require identifiable operators to record each source and stated legal basis for protected model-development material.

    AI providers should maintain proportionate, standardised and machine-readable work-level records of protected material used throughout model development, including material gathered by scraping bots. Operators should be identifiable and record each source and claimed legal basis for use.

  16. Require preservation of provenance information associated with protected material.

    Providers should also preserve copyright- management and provenance information associated with protected material.

  17. Keep providers responsible for establishing provenance and lawful use of third-party training material, rather than shifting responsibility to rightsholders.

    These obligations should extend to datasets obtained from third parties. An AI provider should remain responsible for establishing the provenance and lawful basis of material used in its systems; obtaining training material through an intermediary should not shift that responsibility onto rightsholders.

  18. Allow appropriate protection of genuine confidential information without preventing rightsholders obtaining information needed to identify and enforce uses of their works.

    Appropriate protection may be provided for genuine confidential information, but commercial confidentiality must not prevent rightsholders from obtaining information necessary to identify and enforce uses of their works.

  19. Ensure a provider that fails to comply with mandatory transparency does not benefit from evidentiary uncertainty created by its own non-disclosure.

    A provider that fails to comply with the proposed mandatory transparency should not benefit from evidentiary uncertainty created by its own non-disclosure.

  20. Subject providers offering AI systems in Australia to meaningful transparency and accountability requirements regardless of where model development occurred.

    Providers offering systems in Australia should be subject to meaningful transparency and accountability requirements regardless of where model development occurred.

  21. Require Commonwealth AI suppliers to maintain auditable copyright governance, disclose relevant training practices and demonstrate appropriate licences or another lawful basis where protected material is used.

    Commonwealth procurement should require AI suppliers to maintain auditable copyright-governance processes, disclose relevant training practices and demonstrate appropriate licences or another lawful basis where protected material is used.

  22. Release any proposed AI copyright, licensing or liability model for full public consultation before adoption.

    Any proposed AI copyright, licensing or liability model must be released for full public consultation before adoption.

Read the original

Open in the interactive guide