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 189

National Association for the Visual Arts

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.

Sets out the concerns of visual artists about generative AI, supporting existing copyright protections, voluntary licensing and mandatory training-data transparency, and opposing a text and data mining exception, opt-out models and expanded safe harbours. Also addresses Indigenous Cultural and Intellectual Property, income displacement and Commonwealth procurement.

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. Recommends rejecting a text and data mining exception, opt-out licensing and expanded safe harbours for AI developers.

    NAVA opposes a text and data mining exception, opt-out licensing models and any expansion of safe harbour provisions for AI developers.

  2. Recommends artists' credit, consent and compensation, including knowledge of use, permission or refusal, attribution and payment for authorised use.

    NAVA supports credit, consent and compensation. Artists should be able to know when their work is used, grant or refuse permission, be appropriately credited and receive payment when they authorise its use.

  3. Says use of First Nations material must follow community-determined protocols and free, prior and informed consent, and that a general opt-out is not adequate consent for ICIP.

    The use of First Nations artworks, cultural material and knowledge must follow community- determined protocols and recognise the authority of relevant knowledge holders and communities. This includes free, prior and informed consent, attribution and respect for cultural integrity. A general opt-out mechanism is not an adequate form of consent for ICIP.

  4. Argues overseas training on Australian artists' work followed by Australian model availability leaves artists with limited options for unauthorised overseas use.

    A significant gap arises where Australian artists’ work is used to develop AI models overseas and those models are then made available in Australia. Australian copyright law applies to activities in Australia, leaving artists with limited options where their work has been used overseas without permission or payment.

  5. Argues replacement of early-career commercial work by generative AI removes artists' income and opportunities to establish careers.

    Early-career and lower-value commercial work provides artists with income, experience, professional relationships and opportunities to build a practice. When this work is replaced by generative AI, artists lose income as well as the opportunities through which careers are established.

  6. Supports the Commonwealth procuring AI only from suppliers that have entered licensing arrangements with rightsholders.

    As a co-signatory to the joint rightsholders submission, NAVA supports the Commonwealth procuring AI products and services only from suppliers that have entered licensing arrangements with rightsholders.

Recommendations

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

  1. Keep AI licensing voluntary and reject opt-out approaches and expanded safe harbours for AI developers.

    that licensing should remain voluntary, that opt-out approaches and expanded safe harbour protections for AI developers should be rejected

  2. Include artists' displaced income and professional opportunities in assessments of AI productivity benefits.

    Where work previously undertaken by artists is replaced by generative AI, there is a corresponding loss of income and professional opportunities. These impacts need to form part of any assessment of the productivity benefits of AI.

  3. Develop Australian sovereign AI consistently with copyright and artists' rights, without relying on unauthorised use of artists' work.

    Australian sovereign AI capability should be developed in accordance with Australian copyright law and artists’ rights. Where Australian artworks are used to train or develop AI models, artists should retain control over whether their work is used and be paid where they authorise its use. First Nations cultural material and knowledge require appropriate cultural authority and additional protections. Holding a physical artwork or digital reproduction does not necessarily give a cultural institution or collecting body the copyright required to authorise its use for AI training. Development of Australian AI capability should not rely on the unauthorised use of artists’ work.

  4. Support sector-specific AI tools that protect artists' rights and involve artists and communities in decisions about their work and data, in investment in Australian AI capability.

    Investment in Australian AI capability should support sector-specific tools that protect artists’ rights and involve artists and communities in decisions about how their work and data are used.

  5. Require First Nations material use to follow community protocols, knowledge-holder authority, free, prior and informed consent, attribution and cultural integrity; a general opt-out is inadequate.

    The use of First Nations artworks, cultural material and knowledge must follow community- determined protocols and recognise the authority of relevant knowledge holders and communities. This includes free, prior and informed consent, attribution and respect for cultural integrity. A general opt-out mechanism is not an adequate form of consent for ICIP.

  6. Introduce standalone ICIP legislation and make AI regulation consistent with it, recognising First Nations cultural authority and Indigenous data sovereignty.

    NAVA supports the Australian Government’s commitment to standalone ICIP legislation. AI regulation must be consistent with that legislation and recognise First Nations cultural authority and Indigenous data sovereignty.

  7. Account for AI's effect on artists' incomes, work and career pathways in assessing job creation and displacement.

    The Committee’s assessment of job creation and displacement must account for the effect of AI on artists’ incomes, work and career pathways.

  8. Procure AI products and services only from suppliers that have entered licensing arrangements with rightsholders.

    NAVA supports the Commonwealth procuring AI products and services only from suppliers that have entered licensing arrangements with rightsholders.

  9. Require Commonwealth purchasers to obtain transparency about training copyright material, including whether artworks were used and rightsholders authorised use.

    Commonwealth agencies should also require transparency about the copyright material used to train the systems they purchase. For visual artists, this includes whether artworks were used in training and whether artists or other rightsholders authorised that use.

  10. Ensure agency use of artists' grant and program information in or with AI systems has clear consent, governance and transparency.

    Commonwealth agencies also hold substantial information supplied by artists and arts organisations through grants, applications, acquittals and other programs. The use of this information in or with AI systems requires clear consent, governance and transparency about how artists’ data and intellectual property are used.

  11. Include clear requirements on AI use, copyright, moral rights and ICIP in government arts funding; public money should not support AI products built on artists' work without permission or payment.

    Public money should not support AI products developed using artists’ work without permission or payment. Government funding for the arts should also include clear requirements concerning AI use, copyright, moral rights and ICIP.

  12. Maintain existing copyright protections and do not introduce a text and data mining exception.

    NAVA supports maintaining existing copyright protections and the Australian Government’s commitment not to introduce a text and data mining exception.

  13. Transparency requirements should let artists check whether their work is in training datasets and require developers to disclose the source, how it was obtained and the legal basis for its use.

    Transparency requirements should give artists a practical way to check whether their work has been included in training datasets and require AI developers to disclose the source of training material, how it was obtained and the legal basis for its use.

  14. Give further consideration to AI models developed overseas using Australian artists' work without permission and then made available in Australia.

    NAVA agrees with Copyright Agency that further consideration is needed where AI models developed overseas using Australian artists’ work without permission are then made available in Australia.

  15. Introduce regulatory measures giving artists practical ways to identify whether their work has been used, exercise their rights and seek redress.

    NAVA supports regulatory measures that give artists practical ways to identify whether their work has been used, exercise their rights and seek redress.

  16. Retain artists' ability to decide and negotiate AI training use, including payment; provide knowledge of use, consent or refusal, attribution and payment for authorised use.

    Artists should retain the right to decide whether their work is used to train AI and to negotiate the terms of that use, including payment. NAVA supports credit, consent and compensation. Artists should be able to know when their work is used, grant or refuse permission, be appropriately credited and receive payment when they authorise its use.

  17. Introduce mandatory transparency requirements for companies that compile, create and use AI training datasets, including the sources of material.

    NAVA supports mandatory transparency requirements for companies that compile, create and use AI training datasets, including transparency about the sources from which material has been obtained.

  18. Make licensing arrangements work for individual artists as well as major rightsholders; complexity is no reason to reduce rights.

    For visual artists, any licensing arrangements need to work for individual artists as well as major rightsholders. The so-called “long tail” includes thousands of independent artists whose livelihoods depend on their ability to control, license and earn income from their work. Difficulty licensing large numbers of individual rightsholders is not a reason to reduce those rights.

  19. Ensure licensing arrangements recognise different rights, markets and practices across creative sectors.

    Licensing arrangements also need to recognise the different rights, markets and practices across creative sectors.

  20. Support collective licensing only with voluntary participation and artists retaining control over whether and how work is used.

    Collective licensing can assist with licensing at scale where participation is voluntary and artists retain control over whether and how their work is used.

  21. Invest in licensing infrastructure enabling individual artists and smaller rightsholders to participate in emerging AI licensing markets.

    NAVA also supports investment in licensing infrastructure that enables individual artists and smaller rightsholders to participate in emerging AI licensing markets.

  22. Include effective moral rights protections, transparency about AI-generated images and specific ICIP protections in the regulatory response to AI harms.

    NAVA supports effective moral rights protections, transparency about AI-generated images and specific protections for ICIP as part of the regulatory response to AI harms.

Read the original

Open in the interactive guide