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 207

Association of Artist Managers

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 should keep strong copyright protection against AI training and extend it to artists through consent, remuneration, moral rights and labelling rules. Also seeks algorithmic transparency from streaming and social platforms, and asks that creative IP be treated as sovereign capability. States that AI can create new opportunities for music.

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. States that preserving strong copyright protection is the first and most important objective in the current AI policy environment.

    AAM's position is unambiguous: the preservation of strong copyright protections is the first and most important objective in the current AI policy environment.

  2. Asks for platforms to report aggregate outputs of what their recommender systems served, without disclosing the algorithm itself.

    We are not asking any platform to disclose its algorithm. We are asking for outputs — what the system actually served, in aggregate.

  3. Argues that an arrangement not requiring the rightsholder's consent on the specific works used is not a licence.

    A licence is granted by the person who holds the right, on terms they negotiate, with the ability to refuse. An arrangement that does not require that consent on the specific works being used is not a licence, and should not be described as one.

  4. Argues that copyright protection for rights holders does not equal protection for artists, who often assigned their rights under standard contracts.

    Copyright protection for rights holders is not the same as protection for artists. Where an artist is signed, a record label holds the copyright in a sound recording; the artist performed on it and may have written it, but assigned their rights under standard industry contracts.

  5. Says APRA AMCOS estimates that 23% of music creators’ revenues in Australia and New Zealand will be at risk from generative AI by 2028.

    APRA AMCOS estimates that by 2028, 23% of music creators’ revenues in Australia and New Zealand will be at risk from generative AI.

  6. Recommends that creative intellectual property and workforce be recognised as sovereign capability under the National Interest Framework test of independent Australian control.

    That the Committee recognise Australia's creative intellectual property and creative workforce as sovereign capability, and that the National Interest Framework test of independent Australian control be applied to Australian creative and cultural content on the same terms it is applied to digital infrastructure.

Recommendations

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

  1. Have the Treasurer direct an ACCC Part VIIA inquiry using compulsory powers to establish Australian music streams and plays by source, Australian content within each source, and AI-generated content surfaced.

    That the Treasurer direct the Australian Competition and Consumer Commission to inquire into music streaming services, and the social media platforms where music is increasingly discovered, under Part VIIA of the Competition and Consumer Act 2010 — the mechanism used for its digital platforms inquiries — using its compulsory information-gathering powers to establish, for the Australian market, the share of music streams and plays by source (editorial, algorithmic, radio, autoplay, recommendation feed and user-selected), the representation of Australian content within each, and the share of AI-generated content surfaced.

  2. Require platforms with Australian presence or material Australian revenue to report annually on passive/algorithmic versus active music listening/viewing, Australian content within playlists/autoplay/feeds, and AI-generated content surfaced.

    That Australia legislate standing algorithmic transparency obligations for platforms with an Australian presence or material Australian revenue — digital service providers and social media alike — requiring annual public reporting on the proportion of music listening and viewing attributable to passive or algorithmically-driven modes versus active user selection, the relative representation of Australian content within algorithmic playlists, autoplay and recommendation feeds, and the share of AI-generated content surfaced; and giving regulators and vetted researchers confidential access to the methodology by which content is recommended, surfaced or suppressed, and to the feedback mechanisms by which early engagement data influences downstream recommendation.

  3. Give regulators and vetted researchers confidential access to recommendation, surfacing and suppression methodology and early-engagement feedback mechanisms.

    and giving regulators and vetted researchers confidential access to the methodology by which content is recommended, surfaced or suppressed, and to the feedback mechanisms by which early engagement data influences downstream recommendation.

  4. Preserve strong copyright protections as the primary AI policy objective, with artist protections complementing them.

    AAM's position is unambiguous: the preservation of strong copyright protections is the first and most important objective in the current AI policy environment. Everything else in this section depends on it. These proposals are complementary to the copyright fight — they address what happens to artists once copyright is protected, not instead of protecting it.

  5. Close off settlements, access schemes or permits that give AI companies creative works for centralised payments or limited-party deals without creators’ consent to the specific works used.

    The framework that gives effect to it should also close off arrangements that would achieve a similar result under another name — a "settlement", an "access scheme", or a "permit" — under which AI companies gain access to creative works in exchange for centralised payment, decoupled from the works used –or – for commercial deals with a limited set of parties, without the consent of the people who made the work.

  6. Do not describe access arrangements lacking consent on the specific works used as licences.

    A licence is granted by the person who holds the right, on terms they negotiate, with the ability to refuse. An arrangement that does not require that consent on the specific works being used is not a licence, and should not be described as one.

  7. Introduce a dedicated moral right against unauthorised machine learning, with protections applying where a model trained offshore is deployed in Australia.

    AAM supports Yuvaraj's recommendation to introduce a dedicated moral right against unauthorised machine learning, and endorses the flexible territoriality approach that would apply protections where a model trained offshore is deployed in Australia.

  8. Examine Yuvaraj’s moral-rights research closely.

    Yuvaraj is on the Attorney-General’s Copyright and Artificial Intelligence Reference Group and his research should be examined closely.

  9. Protect active, specific artist and songwriter consent before AI use of their works, voice, performance, likeness or creative identity.

    Consent & Control. “Artists and songwriters must actively and specifically consent before their works, voice, performance, likeness or creative identity is used in connection with AI.”

  10. Require purpose-specific requests for AI use.

    Clarity & Transparency. Requests must be purpose-specific.

  11. Protect creators’ moral, neighbouring, image and personality rights from label or publisher licensing without clear authority, consent and accountability.

    Voice, likeness and creative identity need particular attention. The cause is explicit that artists and songwriters remain the primary holders of many of the rights at stake, “including moral, neighbouring, image and personality rights”, and that these rights “are not label or publishers assets to be licensed without clear authority, consent and accountability.”

  12. Require fresh, express artist consent for training on catalogues covered by contracts signed before AI training existed; those contracts must not alone authorise training.

    A rule that contracts signed before AI training existed do not, on their own, authorise it. Using those catalogues for training should require the artist's fresh, express consent.

  13. Ensure participating artists and songwriters receive fair, meaningful remuneration, are consulted and know the revenue shares going to creators, labels/publishers and AI companies.

    As the “no consent, no deal” cause puts it: “Where artists and songwriters choose to participate in AI-related uses, they must receive fair and meaningful remuneration. They must share in the value created by their works, and it cannot be considered a general label asset. Artists and songwriters must be consulted and it must be clear which percentage of revenue goes to the creator, to the label and to the AI company.” The same applies to publishers.

  14. Where causal attribution to training works is technically feasible, use a collecting society Equitable Remuneration model delivering generation and tracking royalties.

    Where causal attribution is technically feasible ('Computational Copyright'), a collecting society Equitable Remuneration (ER) model is appropriate, delivering generation royalties (at the moment of creation) and tracking royalties (ongoing performance, mechanical, and sync revenues for the life of the new work).

  15. For catalogue training or any one-time AI training fee or guarantee paid to a label, require a statutory minimum percentage paid directly to featured performers/songwriters at contract level, free of recoupment.

    Instrument two is for where a model is trained on a catalogue and the commercial event is complete, or when a label receives any one-time training fee or guarantee from the AI company, a statutory pass-through obligation should exist. A defined minimum percentage must flow directly to the featured performer or songwriter, at the contract level, as a matter of law, free of recoupment.

  16. Strengthen moral rights in Australian law, specifically the right of integrity, so they extend to unauthorised use of works to train AI models.

    further support the need for moral rights to be strengthened in Australian law, specifically the right of integrity, so that it can be applied and extended to the unauthorised use of works to train AI models.

  17. Require mandatory AI labelling and watermarking disclosure now, even before technical watermarking can fully verify it.

    Mandatory AI labelling and watermarking is the evidentiary infrastructure on which every other creator protection depends. Without it, consent violations are invisible, pass-through obligations cannot be audited, and moral rights claims cannot be brought. Australia requiring mandatory disclosure now — even before technical watermarking can fully verify it — establishes that infrastructure before the scale of AI-generated music makes retroactive evidence-gathering impossible.

  18. Implement measures ensuring artist control, protection and remuneration across AI-related regulation, such as AAM's four-part framework.

    Implement measures that ensure the control, protection and remuneration of the artist at all levels of AI-related regulation, such as the AAM's proposed four-part framework.

  19. Recognise creative IP and creative workforce as sovereign capability, applying the independent Australian control test to creative content.

    That the Committee recognise Australia's creative intellectual property and creative workforce as sovereign capability, and that the National Interest Framework test of independent Australian control be applied to Australian creative and cultural content on the same terms it is applied to digital infrastructure.

Read the original

Open in the interactive guide