AI and Copyright

Authors: Martin Lee, Anna Scotland, Anna Feng, Minami Graham and Benjamin Hung

---

EXECUTIVE SUMMARY

In December 2025, the Productivity Commission released their report Harnessing data and digital technology (‘PC Report’).¹ The PC Report arrives in the context of the increasing presence of artificial intelligence (‘Artificial Intelligence’) across creative industries. The Productivity Commission considered several significant areas for policy reform to boost Australia’s productivity growth using digital technology, including copyright in the context of artificial intelligence (‘Artificial Intelligence’).² 

The Productivity Commission noted in the PC Report that copyright laws provide incentives for creators to produce content by restricting the rights of others to copy and profit from their work.³ The PC Report also acknowledged that AI models are built on large data sets, which are sourced by copying digital media.⁴ In order to train an AI model on these data sources, an AI developer must obtain permission in the form of a licence from the creators of the original content.⁵ Copyright can therefore be a source of income for creators by agreeing to give others a licence to copy their work. However, this has not stopped many AI developers from using vast amounts of copyrighted material without consent or compensation in order to train their models. In response to this novel domain of AI copyright licensing, the PC Report recommended a ‘wait and monitor’ approach over the next 3 years to ascertain whether policy intervention is required and if so, how it should be applied.⁶ 

This paper is purposed to examine whether Australia’s current direct copyright licencing model effectively protects the copyrights of creators in the context of AI model training and licensing. This paper is accordingly structured as follows. This section outlines the original methodology of the research, and how the paper's scope changed to reflect the private nature of the licencing market. Part I sets out findings regarding current licencing markets across the three main creative industries of music and audio, literature, and photography and film. Part II presents a projection of the future of these licencing markets based on the trends identified in Part I.

SCOPE

This paper seeks to examine the licencing market across three categories of creative content in relation to AI model training:

1. Music and audio;

2. Written works, including literature and journalism; and

3. Photography and cinematography.

While this paper identifies some developments in licencing markets across these creative industries, it identifies two significant issues of transparency in analysing the current practical benefits of current copyright laws in Australia for creators. Firstly, this paper identifies a general hesitancy across creative industries to engage with AI developers due to a lack of transparency from AI developers about their products. Secondly, as licencing deals between creators, or those that manage creator content such as music studios or book publishers, are private contracts, public information or transparency on the current licencing market is limited. This paper notes that future research on the licencing market between Australian creators and AI developers may need to acknowledge these issues of transparency.  

This paper is based on information available prior to 8 July 2026 and should not be relied on as reflecting any developments after this date.

METHODOLOGY

Adopting a fact-based approach based on publicly available information, this report originally sought to answer the following three questions regarding the three categories of creative content:

1. What licensing deals have already taken place?

2. Approximately what percentage of creators’ copyright is currently covered by licensing agreements with AI companies?

3. At the current rate, what percentage of creators’ copyright will likely be covered by June 2026, the end of 2026 and the end of 2028?

Due to limited public information on licencing deals, this paper was restructured to firstly analyse the current market of licensing and then project the future of market based on trends identified, moving away from the goal of providing a discrete figure of how much copyrighted work will be covered under AI licensing deals.

The methodologies for each category of creative content are laid out below.

1: MUSIC AND AUDIO

The two main relevant forms of copyright are:

Musical works: while not explicitly defined in the Copyright Act 1968 (Cth), refer to the underlying composition (i.e. the sounds/notes/melodies that can be fixed in a form).

Sound recordings: defined in s 10 of the Copyright Act 1968 (Cth) as a collection of sounds captured on a record (including CD, MP3, vinyl, tape, cassette, etc).

Given the global and highly consolidated nature of the music industry, we initially focused on  the three largest record labels by market share (commonly known as the ‘Big 3’) — Universal Music Group, Sony Music Entertainment and Warner Music Group — and their respective highly publicised settlements and agreements (or lack thereof) with the two largest AI music generation platforms by user base — Suno and Udio.

During the course of our research, we located the Copyright Alliance’s website,⁷ which helpfully compiled an expansive list of publicly announced licensing agreements between music and audio rights holders and AI developers. This list formed the basis of our research, and from here we attempted to identify the specific terms of each agreement.

A key limitation was the opaque nature of many of the licensing agreements, which meant that we had little to analyse other than the rights holder/AI developer’s press release announcing the deal, and speculative reports by journalists or purported industry insiders.

2: WRITTEN WORKS

Pursuant to s 10 of the Copyright Act 1968 (Cth), literary works are defined to include:

1. a table, or compilation, expressed in words, figures or symbols; and

2. a computer program or compilation of computer programs 

To conduct our research, we divided literary works into two categories: news media corporations and major book publishers. We then selected the major players in each of these industries to concentrate our analysis. For news media, these included the public broadcasters ABC and SBS, the commercial platforms News Corp, Nine Entertainment and Seven West Media, and prominent newspaper Guardian Australia. For the publishing industry, we selected the ‘Big Five’ – Penguin Random House, Hachette Australia, HarperCollins Publishers, Pac Macmillan and Simon & Schuster. We also included Black Inc Books, an independent Australian publisher. 

We conducted research on the current state of licensing deals with AI companies in each of these two industries. Our research included examining each industry player’s current position on AI, current or purported licensing deals, the monetary success of these deals, the position of AI companies entering into deals and, critically, the response from journalists and authors. We were limited in our fact-finding analysis by the lack of publicly available information outlining any existing or purported current deals.

3: PHOTOGRAPHY AND CINEMATOGRAPHY

Under s 10 of the Copyright Act 1968 (Cth), a ‘photograph’ is considered as a form of ‘artistic work’,⁸  whereas ‘cinematography’ is separately defined under ‘cinematograph film’.⁹

For photography and short-form cinematography, our research commenced by identifying voluntary licensing arrangements and the reasons behind their limited adoption within this sector. We then evaluated the common modes of production and publication, and analysed licensing clauses contained in the unilateral standard user agreements across various service providers (including editing software, social media and stock platforms), and whether opt-out option is provided for AI training.

In addressing the research questions, our analysis focused on the hyper-fragmented nature of this sector, and the ongoing limitations and risks affecting the individual artists. In light of these findings, we recommend that a centralised licensing framework alone may not fully address the persistent concerns within this sector.

As outlined above, a ‘cinematograph film’ is recognised as a category of copyright subject matter under s 10 of the Copyright Act 1968 (Cth).¹⁰ 

For film, our research commenced by identifying publicly disclosed licensing agreements between AI developers and film studios. However, given the limited number of publicly identifiable licensing agreements, our research subsequently examined the factors contributing to their limited adoption and analysed the current state of AI licensing within the global film industry. This research included reviewing publicly announced commercial licensing partnerships, litigation concerning the unauthorised use of copyrighted films for AI model training, existing industry protections for rightsholders, and industry statements.

In addressing the research questions, we also analysed major studios’ publicly identifiable positions towards AI licensing and, together with film industry market share data, adopted a scenario-based approach to develop projections of the future proportion of films likely to be covered by licensing agreements. 

We acknowledge our limited ability to quantify the exact proportion of films covered by licensing agreements due to the lack of publicly available information, as many agreements remain confidential or have not yet been publicly disclosed. Accordingly, our projections are illustrative rather than predictive and are derived from the information available at the time of writing.

---

1: ANALYSIS OF CURRENT LICENSING MARKET

A: MUSIC AND AUDIO

1: THREE IDENTIFIABLE AREAS

The development of licensing agreements between music/audio copyright holders and AI companies can be split into three identifiable time periods. First, the ‘Wild West’ (until Q2 2024), followed by litigation (Q3 2024 to Q3 2025), and then settlements and formalised licensing (Q4 2025 to present).

A): The Wild West

Suno launched on Discord in July 2023,¹¹ and launched its web platform (suno.com) in December 2023.¹² Udio launched publicly in April 2024 with strong financial backing.¹³ While AI music generation had existed previously, these two platforms made it widely accessible to the public for the first time and became highly popular with users. The ‘Wild West’ was characterised by legal uncertainty, with GenAI music platforms training their models on vast amounts of illegally obtained material and openly admitting to such activity, but claiming this to be ‘fair use’.¹⁴ No major licensing deals took place in this era, due to the perceived (or purported) lack of necessity on the part of AI companies to seek permission from or compensate music and audio copyright holders.

B): Litigation

The shift from the ‘Wild West’ to the era of litigation was marked by the Recording Industry Association of America (RIAA) suing Suno and Udio on behalf of Universal Music Group (UMG), Sony Music Entertainment (SME) and Warner Music Group (WMG) (‘The Big 3’) for mass infringement of copyrighted works in June 2024.¹⁵ Digital Service Providers (DSPs) such as Deezer and Spotify began flagging and removing AI-generated tracks from their platforms.¹⁶ Meanwhile, Krystle Delgado (also known as Top Music Attorney, a prolific music lawyer) launched a class action lawsuit against Suno and Udio on behalf of independent artists.¹⁷

C): Settlements and Formalised Licensing

In mid 2025, the era of settlements and formalised licensing began, marked by Universal being the first of The Big 3 to settle with Udio in October 2025, and Warner settling with Suno and Udio in November 2025. This was followed by a significant and rapid uptick in formalised agreements between copyright holders and AI companies. Following the settlement, Udio disabled users’ downloading of outputs generated on its platform, angering its users.¹⁸ This highlights the increased compliance with the demands of rightsholders by AI companies in this era, as well as a shift toward ‘walled-garden’ models more generally.

2: WHAT LICENSING DEALS HAVE ALREADY TAKEN PLACE?

Below is a list of publicly reported licensing deals between music and audio copyright holders and AI companies:

Table 1: Publicly reported music/audio licensing deals with AI companies

3: WHAT PERCENTAGE OF MUSIC AND AUDIO IS CURRENTLY COVERED BY LICENSING AGREEMENTS WITH AI COMPANIES?

The Big 3 (Universal, Sony, Warner) make up a majority of rights to recorded music (~85%).⁴⁴ Together with Merlin, they make up ‘The Big 4’ and cover the vast majority of recorded music. The remainder, consisting of smaller independent rightsholders, may choose to opt in to having their work be included in AI training models via their distributor (e.g. Symphonic Music), if the distributor has formed a partnership with an AI company (e.g. Musical AI).

Indeed, though the specific details of every deal may not be available, there are identifiable emerging models of compensation. These include:

1. Lump sum buyouts: one-time payments whereby the rightsholder provides unrestricted access to their catalogue or archives. This provides immediate liquidity for rightsholders, but no future compensation should the AI platform become more valuable. The attitude of Suno and Udio around training data contrasts that of traditional tech companies such as YouTube, which has reportedly offered lump sums to labels in return for permission to train its models on their catalog.⁴⁵

2. Ongoing revenue sharing: mirrors the traditional streaming model of platforms like Youtube or Spotify, whereby rightsholders receive a percentage from a pool of the AI platform’s subscription or ad revenue based on how often their content was used for generation. Such attribution can be difficult to prove compared to a simple stream count.

3. Pay-per-value (attribution modelling): determine exactly how much a specific human work contributed to or influenced a specific AI-generated track. This shifts the onus from mere streams and downloads and instead credits rightsholders in terms of actual influence. AI companies like Sureel have offered this model.⁴⁶

4. Hybrid tech credits: a barter-style compensation whereby the rightsholder receives API credits instead of or alongside cash, allowing them to then propel their own AI and digital infrastructure forward. While this may work well for large enterprises, it does not always result in direct royalties for creators.

B: WRITTEN WORKS

1: WHAT LICENSING DEALS HAVE ALREADY TAKEN PLACE?

Australian copyright law grants rightsholders the exclusive right to reproduce their copyrighted works.⁴⁷ AI developers must obtain permission from the rightsholder to use these works to train AI models or risk infringing the rights of the copyright holder. Permission is granted in the form of a direct license between the AI developer and the rightsholder and is typically obtained by compensating the rightsholder. Most direct licensing deals have engaged with ‘high value’ works from established copyright holders, including news media corporations and major book publishers. The Productivity Commission categorised these deals as a ‘Category 2’ copyright status out of 3 due to the easily identifiable rightsholder and the straightforward process in negotiating a license as licensing these works for purposes outside of AI training already exist.⁴⁸ 

A): News Media Corporations

News media corporations are actively protecting their journalism through licensing deals with AI developers and pursuing litigation to demand compensation for the “scraping” of their content to train AI models.

i): News Corp and OpenAI Partnership (2024)

The News Corp and OpenAI partnership of 2024 is a significant, multi-year licensing deal allowing Open AI to access current and archived content from over fifteen News Corp publications.⁴⁹ These publications include, The Wall Street Journal, Barron’s, MarketWatch, Investor’s Business Daily, Finance News Network (FN), New York Post, The Times, The Sunday Times, The Sun, The Australian, news.com.au, The Daily Telegraph, The Courier Mail, The Advertiser and Herald Sun. The partnership grants OpenAI permission to generate content from these publications in response to user questions as well as use News Corp’s content to enhance its AI products. News Corp will also provide “journalist expertise” to the AI developer that will allow OpenAI to commit to maintaining high standards of journalism across its AI offerings.⁵⁰ The partnership is alleged to be worth more than USD $250 million (AUD $378 million) for the five-year partnership.⁵¹

News Corp’s 2026 partnership with Meta Platforms is another significant deal for the news media industry.⁵² Concerning content from its US and UK platforms, News Corp granted Meta permission to use its content to train AI models. Lasting for at least three years, News Corp is expected to earn up to USD $50 million per year from the partnership.

ii): Guardian Media Group and OpenAI Partnership (2025)

The Guardian Media Group entered into a partnership with OpenAI in 2025, providing editorial content for use in ChatGPT’s outputs.⁵³ Users of ChatGPT can access and receive summaries of Guardian reporting and archive journalism with the aim of expanding the Guardian’s reach to ChatGPT’s over 900 million users.⁵⁴ The Guardian will also implement ChatGPT Enterprise, an advanced version of ChatGPT with higher privacy and security tools, across its business.

iii): Nine Entertainment

Nine Entertainment has engaged two licensing deals with “key domestic corporates”, providing its content to train in-house large language models (LLMs).⁵⁵ As part of the News Media Bargaining Code introduced in 2021, Nine additionally contracted with Google to provide its content in YouTube and search engine results, however the contract will expire this year.⁵⁶ Last year, Nine accused OpenAI, Perplexity and other major AI developers of illegally “scraping” its content,⁵⁷ recording AI developers engaged with its websites almost “10 times every second” in June of 2025 with OpenAI accounting for “several of those 10”.⁵⁸ Nine CEO Matt Stanton has been vocal in his advocacy for receiving compensation from the illegal scraping, calling on the Albanese government to prioritise assisting the news media industry enter into licensing deals.⁵⁹ 

iv): ABC, SBS, Seven West Media and Australian Community Media (ACM)

ABC, SBS, Seven and ACM allegedly formed an alliance in late 2024 to negotiate compensation from AI developers using their respective content to train AI systems, however publicly available information is limited.⁶⁰ The ABC has also directly accused AI developers of scraping its content despite blocking ChatGPT, Perplexity and Claude,⁶¹ blocking OpenAI’s GPTBot in 2023, a web crawler able to scan webpages for content to improve its AI models, alongside the New York Times, CNN and Reuters.⁶² 

B): Publishing

Few major book publishers have entered into licensing deals with AI developers and many are actively attempting to prevent the use of copyrighted works to train AI models and most are involved in copyright advocacy. Those that have licensed their authors’ works, typically with unnamed AI developers, have been met with extensive backlash from authors and literary organisations.

i): HarperCollins Publishers and Unnamed AI Developer (2024)

In 2024, HarperCollins entered into an agreement with an unnamed AI company, the first licensing deal to be made by a major Australian publisher with an AI developer. The agreement allows for “limited use of nonfiction blacklist titles for training AI models” with the permission of the author and involves over 150 Australian authors.⁶³ The deal takes the form of an opt-in contract, offering authors USD $2,500 (AUD $3,500) per title for a three-year license.⁶⁴ The amount of compensation corresponds to a 50/50 split common in AI licensing deals, the AI developer offering HarperCollins USD $5,000 which the publisher then evenly halves between the author and themselves. The Australian Society of Authors (ASA) advocates for a fair compensation of a 75/25 split, representing the US Author’s Guild guidance on a fair split where 75% is provided to the author as it is the work of the author that is of “the most value” to AI developers.⁶⁵

The terms of the contract are not negotiable,⁶⁶ however HarperCollins has negotiated four ‘guardrails’ the AI developer must account for, according to the ASA, of which include:

1. Limit the AI model’s outputs to no more than 200 consecutive words and/ or five percent of a book’s text across multiple outputs in a user session

2. Prohibit commercial users from attempt to infringe the copyright of the books, and must monitor those users who do infringe copyright

3. Pledge not to scrape data from an agreed list of pirate websites

4. Promptly address any breaches of copyright⁶⁷ 

ii): Penguin Random House

Penguin Random House is an active advocate for human creativity and upholding current copyright laws. The publisher has not engaged in licensing its authors' works to AI developers. It supports the developing licensing market,⁶⁸ and has joined the Association of American Publishers (AAP) alongside a coalition of more than 10,000 creators and cultural institutions condemning the unlicensed use of copyrighted works in AI training.⁶⁹ In 2024, the publisher amended its copyright wording across its global imprints to read, “No part of this book may be used or reproduced in any manner for the purpose of training artificial intelligence technologies or systems”.⁷⁰

iii): Hachette Australia

Hachette Australia is “opposed to ‘machine creativity’”,⁷¹ and has not entered into any licensing deals with AI developers, instead advocating for creative rights for copyright holders. Its US imprint is heavily involved in a current class action case against Meta Platforms for the alleged use of copyrighted content in training AI models.⁷² Additionally, the publisher is attempting to join a class action case against Google in the US as a representative for publishers whose copyright has been infringed by Google, however Google has opposed their motion to intervene in the case.⁷³ 

iv): Pan Macmillan Australia

Pan Macmillan, the international publishing house of which Pan Macmillan Australia is an imprint, is involved in advocacy concerning the ethical use of AI and the implementation of appropriate legislative provisions that protect authors and illustrators.⁷⁴ The publisher has not engaged in any publicly known licensing deals yet and, as with Hachette, its US imprint is suing Meta Platforms in a class action case over the use of copyrighted works in AI training.⁷⁵ The publisher’s UK imprint is also a member of the Publisher’s Association AI Taskforce and the Creative Rights in AI Coalition (CRAIC) which launched the ‘Make it Fair’ campaign,⁷⁶ advocating for the UK government to uphold its current copyright laws in the face of legislative change.⁷⁷

v): Simon & Shuster

Meta Platforms considered acquiring the publishing house Simon & Schuster in 2023, according to The New York Times.⁷⁸ The acquisition would have allowed Meta to access copyrighted books to train its AI models. According to The Times’ records of internal meetings at Meta, Ahmad Al-Dahle, Meta’s vice president of generative AI, discussed concerns the company had exhausted nearly every book, poem and essay written in English on the internet and was in need of new material for its models.⁷⁹ Indeed, it was reported Meta executives said negotiating licenses with rightsholders “would take too long”,⁸⁰ a sheer representation of the tech company’s respect for intellectual property.

vi): Black Inc Books

Independent, Melbourne-based publisher Black Inc Books provided authors with an opt-in contract to partner with unknown, “reputable” AI developers. Black Inc asked authors to grant the publisher the right to reproduce, use, adapt and exploit their works for the development of software training, including AI models.⁸¹ The ASA argued Black Inc did not provide adequate information for authors to evaluate the contract nor time to seek legal advice, saying “blanket permission for all future licensing – particularly with only days to sign – is unnecessary and unfair”.⁸² Additionally, the ASA again reiterated the 50/50 split for compensation did not represent fair compensation, instead supporting the 75/25 split put forward by the US Authors Guild.⁸³ 

2: WHAT PERCENTAGE OF LITERARY WORKS ARE CURRENTLY COVERED BY LICENSING AGREEMENTS WITH AI COMPANIES?

It is difficult to determine the percentage of literary works covered by licensing deals with AI developers, largely due to the lack of publicly known information surrounding current and purported deals. The following analysis quantifies the above licensing deals engaged by news media organisations and major book publishers. Importantly, the analysis is not an exhaustive collation of literary works currently covered by licensing deals as the focus is on major industry players rather than independent and smaller organisations.

Regarding news media corporations, our findings show that out of the 7 corporations analysed, only 3 have entered into licensing deals with AI developers, and out of those 3, 2 with OpenAI (Table 1.1). Additionally, 5 out of the 7 are actively involved in negotiations for compensation and have publicly accused AI developers of content scraping (Table 1.2). Nine in particular is calling on the Albanese government to assist the news media industry to secure licensing deals with AI developers.⁸⁴ 

Table 2: Licensing deals between major news corporations and AI developers

Table 3: Advocacy by major news media organisations against AI content 'scraping'

For the publishing industry, only 2 out of the 6 examined have entered into AI licensing deals (Table 2.1). Meta Platforms considered acquiring Simon & Schuster in 2023, however this never materialised. Many publishers are involved in advocacy with 2 out of the 6 actively involved in advocacy groups (Table 2.2). Additionally, 3 are currently involved in court litigation against AI developers across their global imprints (Table 2.2).

Table 4: Licensing deals between major book publishers and AI developers

Table 5: Advocacy by major book publishers against AI content 'scraping'

C: PHOTOGRAPHY AND CINEMATOGRAPHY

1: WHAT PHOTOGRAPHY LICENSING DEALS HAVE ALREADY TAKEN PLACE?

Australian Associated Press, one of the largest suppliers to Australian news, along with its subsidiaries (e.g. AAP Photos) entered into a licensing agreement with Google Gemini in 2025.⁸⁵ Similarly, among major global photo/video stock platforms, Getty Images and Shutterstock both have entered into partnerships with one or more AI companies. Some of the existing licensing partnerships are summarised in the following timeline diagram.⁸⁶

The landmark copyright cases include:

1. Getty Images v Stability AI – in both the UK and US.⁸⁷ The UK case was filed in January 2023, judgment delivered in November 2025 against Getty Images;⁸⁸ while the US case filed in February 2023 remains ongoing.

2. Ongoing class action (by 10 famous visual artists) against Stability AI, DeviantArt, Midjourney, Runway AI, filed in January 2023.⁸⁹

3. Class action against CapCut/ByteDance, i.e. Rodriguez v ByteDance, filed in July 2023.⁹⁰

4. Ongoing class action (by content creators) against Snap, filed in October 2023.⁹¹

5. Austin Beaulier (an American 3D artist) filed action against Meta, Nvidia, Roblox in early 2026.⁹²

2: WHAT PERCENTAGE OF PHOTOGRAPHY WORKS ARE CURRENTLY COVERED BY LICENSING AGREEMENTS WITH AI COMPANIES?

Beside the abovementioned voluntary organisational consents (indirectly given by individual photographers), we have identified two channels (editing and social media platforms with embedded AI functionalities), which may give rise to copyright concerns. This could be through web-scrapping or imbalanced user agreements with broad licensing clauses that favour platform/channel operators, whilst undermining photographers’ explicit consents. The nature of this sector appears to be highly fragmented,⁹³ and continues to evolve in response to demands of the growing digital economy. Certain aspects of photographers’ creative processes remain highly under-regulated. We will review three identified areas below.

A): Digital Stock Platforms

One survey found that 27% participating photographers use digital platforms to generate income, whilst 14% amongst them utilise stock platforms.⁹⁴ A separate survey also demonstrated that a marginal number of photographers engage with stock libraries.⁹⁵ These platforms facilitate the storage and monetisation of their work, generating passive income through bidding and sublicensing. Among these, stock platforms (e.g. Getty Images) are frequently used by marginalised or vulnerable artists as entry points.

The Website Terms of Getty Images (in the contributor community site) specify that the intellectual properties are ‘owned by Getty Images, its licensor, or its third-party image partners’.⁹⁶ The contributors are refrained from a range of acts which may constitute copyright infringement, including data mining.⁹⁷ Most importantly, contributors grant ‘Getty Images and its partners with a non-exclusive, non-revocable, sublicensable, worldwide, royalty-free licence to use any inputs...[the contributor] provide[s] on the Site…’.⁹⁸ In return, eligible contributors receive monthly performance based royalty (e.g. licensing transactions).⁹⁹ Additionally, Getty Images’ built-in AI function supports commercial use, that its training data is derived ‘[e]xclusively [from]… licensed content, including our creative library and data’.¹⁰⁰

B): Editing Platforms

Content editing remains an essential step of the creative process for photographers. As such, photographers/creators are inevitably required to adhere to the licensing clause set out in the user agreements of the editing platforms (e.g. Canva, Adobe, ByteDance, etc.).

Under Canva’s Terms of Use, clause 4a specifies that the users retain rights on the content they legitimately own and upload, whilst granting Canva a broad licence to ‘display, host, copy, store and use your User Content to provide the Service to you, including to keep the Service safe and secure…’.¹⁰¹ Additionally, where users share contents publicly, the licence extends to ‘a perpetual, royalty-free, sublicensable, license to display, host, copy, store and use your User Content to the extent necessary to continue to make that Design available’;¹⁰² the impacted users may choose to update their preference via privacy settings for the participation in AI data training.¹⁰³

Similarly, Adobe also provides users with an ‘opt-out’ option for data training.¹⁰⁴ The platform ensures users that their built-in AI are trained on Adobe stock contents compliant with licensing and copyright that allows ‘confident’ publication of its outputs.¹⁰⁵ ByteDance, on the other hand, has attracted criticisms and class actions for its data training transparency.¹⁰⁶

C): Social Media

Social media has become increasingly significant in the digital era, enabling photographers to build and connect with global communities, whilst meeting their financial needs through self-marketing and commercial works for other businesses.¹⁰⁷

Social media, as the most targeted channels for web-scraping,¹⁰⁸ has evidently proven its popularity over the years.¹⁰⁹ Additionally, Australian businesses have reported to be placing significant focus on social media marketing due to proven results.¹¹⁰  Small businesses have reported to spend up to 30% of marketing budget towards social media contents.¹¹¹

According to the licensing clause sets out in the Terms of Service of Meta platform (Facebook), whilst the users retain copyright over their content, by sharing a content, the user grants Meta a ‘non-exclusive, transferable, sub-licensable, royalty-free, and worldwide license to host, use, distribute, modify, run, copy, publicly perform or display, translate, and create derivative works of your content’.¹¹² which allows them to share with other ‘Meta Products or service providers’.¹¹³ Instagram’s licensing clause is similarly worded.¹¹⁴ Additionally, Meta trains data from publicly available and licensed information, including those that are publicly ‘shared on Meta Products’,¹¹⁵ such as Facebook and Instagram.¹¹⁶

3: AT THE CURRENT RATE, WHAT PERCENTAGE WILL LIKELY BE COVERED THIS YEAR OR BY 2028?

Aside the abovementioned voluntary consent under existing licensing arrangements with AI companies (e.g. through AAP Photos and digital stock platforms), organisations such as the Copyright Agency indicated their willingness to ‘continue to pursue opportunities for licensing for AI-related activity in Australia…[and] have approached AI developers in connection with a range of projects to discuss their licensing needs’.¹¹⁷ The Copyright Agency represents approximately 40,000 members across various professions, including photographers.¹¹⁸

However, there are currently insufficient data available to determine the likely growth in voluntary consent in this sector, despite the likely growth in the stock photography market.¹¹⁹ Moreover, it remains unclear whether the photographers are, or would be, fairly compensated under those licensing arrangements.¹²⁰ Therefore, any increase in the number of licensing agreements would not explicitly translate into an increase in individual photographer’s consent for AI data training.¹²¹ Further, the social media platforms with broad licensing terms continue to present significant risks, and leave artists limited options to protect their works.¹²² 

4: WHAT CINEMATOGRAPHY LICENSING DEALS HAVE ALREADY TAKEN PLACE WITH AI DEVELOPERS?

A limited number of licensing agreements have emerged between AI developers and film studios, typically only in the form of controlled, provenance-based partnerships. This is illustrated by the Lionsgate—Runway collaboration (18 September, 2024), where Runway trained a custom AI model on Lionsgate’s proprietary film catalogue for Lionsgate’s own internal production use.¹²³ This agreement portrays the emergence of commercial licensing within the film sector where there is proper gated access, enforced provenance, and consent. 

However, there remains a general lack of AI licensing deals amongst the major Hollywood studios. Instead, the ‘Big Five’ studios are engaging in litigation against AI developers, as successful injunctions and examination into AI training systems secures future licensing leverage.¹²⁴ This is evident in the ongoing 2025 lawsuit brought by Disney, Universal Pictures, and Warner Bros. against Midjourney for alleged direct and secondary copyright infringement.¹²⁵ Plaintiffs allege that MidJourney scraped their copyrighted works to train its subscription-based AI image generation service without the studios’ consent.¹²⁶ MidJourney responded to these claims on 6 August 2025 by arguing that its AI model operated without memory, and simply generated images through recognition of complex statistical patterns between visual features and user-prompted instructions.¹²⁷ MidJourney reportedly generated $300 million USD through paid subscriptions.¹²⁸ This lawsuit has the potential to become a landmark decision by clarifying whether the use of copyrighted works to train AI models constitutes transformative ‘fair use’ under §107 of the U.S. Copyright Act of 1976,¹²⁹ or amounts to copyright infringement under U.S. copyright law. Given the dominant role of U.S. studios in the global film industry, the decision may have significant implications for future commercial licensing practices. Additionally, the MidJourney dispute highlights the commercial costs arising from the absence of a clear licensing framework protecting rightsholders, whereas the Lionsgate partnership captures the potential financial upside in opt-in, compensated AI licensing arrangements.

The scarcity of licensing agreements within the film industry can be attributed to several factors. Firstly, the scarcity of AI-copyright specific precedent. As potential landmark cases are still ongoing (Disney, Universal, Warner Bros. v Midjourney; Disney, Universal, Warner Bros. v MiniMax),¹³⁰ uncertainty persists regarding future AI-training restrictions and commercial licensing standards. The overall limited engagement from the major studios in licensing agreements, combined with lack of authoritative precedent, create legal uncertainty and a barrier to innovation.¹³¹

Secondly, the inherent complexities of film IP may present various practical difficulties. According to the World Intellectual Property Organisation (WIPO), a film is a collection of copyrights involving various rightsholders, including musicians, actors, set designers, and screenwriters.¹³² The film industry relies on the comprehensive ‘Chain of Title’ method of collective documentation, which proves that all copyrighted elements within the work are licensed and assigned.¹³³ AI training departs from this traditional model, as there is no transparent rights tracking or ‘Chain of Title’ process in the ingestion of films for LLM training.¹³⁴ It is uncertain whether the “wait and see” approach will allow the film industry to naturally navigate and establish licensing norms despite these challenges. 

The ‘Big five’ studios may also have strategic incentives to delay widespread licensing development in order to preserve their market position within the industry,¹³⁵ especially with filmmakers possibly saving up to 50% on big-budget films by using AI for visual effects.¹³⁶

Additionally, there is pushback from some industry workers against the unauthorised use of their works for AI training, such as their voice,¹³⁷ and a general concern by directors on preserving the art and humanness of film in the wake of AI.¹³⁸

One example of established protections against the unauthorised use of film in AI training is the SAG-AFTRA 2023 TV/Theatrical Agreement with the Alliance of Motion Pictures and Television Producers (AMPTP).¹³⁹ Hollywood’s largest labour union for media professionals secured unprecedented protections surrounding AI, including consent requirements for digital replicas, compensation for AI-enhanced productions, and transparency obligations.¹⁴⁰

Despite limited external licensing, there is overall widespread internal adoption and development of AI tools within film studios.¹⁴¹ SAG-AFTRA’s advocacy for AI safeguards and improved residuals may also indicate a call for controlled, provenance-based AI licensing deals that protects rightsholders in the future.¹⁴² 

Overall, the film industry remains in a period of transitional tension regarding AI licensing. The Copyright Alliance reports more than fifty AI-related copyright lawsuits within the past several years,¹⁴³ suggesting that litigation has become the primary mechanism through which rightsholders are seeking to protect their IP. Licensing may become an emerging strategy for studios to protect their IP by allowing AI developers lawful access to their IP while ensuring remuneration for rightsholders. AI developers will also seek licensing as high-quality film footage as training data serves as a major competitive differentiator.¹⁴⁴ However, uncertainty persists due to lack of precedent decisions, limited engagement from major Hollywood studios, industry pushback, and uncertainties around provenance and the AI training process. A compulsory licensing scheme for AI training on copyrighted works may ease uncertainties by establishing baseline requirements for transparency, compensation, and legality in the licensing space, rather than leaving industries to organically navigate commercial licensing standards through lengthy and costly litigation.¹⁴⁵ 

5: WHAT PERCENTAGE OF THEATRICAL FILMS ARE CURRENTLY COVERED BY LICENSING AGREEMENTS WITH AI COMPANIES?

Determining the percentage of films currently covered by AI licensing agreements presents difficulties as we are limited to publicly disclosed information, and many agreements remain confidential. Accordingly, the following analysis relies on publicly stated positions of major film studios together with publicly available evidence of AI partnerships, litigation, and licensing activity to estimate the current and potential future extent of licensing.¹⁴⁶ As major film studios are based in the U.S., there was ultimately a focus on U.S. studios, as their stances on AI licensing has the potential to capture the industry’s position and proportion of films covered. 

Ten film studios represented within the top 60 highest-grossing films at the 2025 Australian box office were categorised according to their publicly identifiable position on AI licensing. This list contained the ‘Big Five’ in film—Walt Disney, Universal Pictures, Warner Bros. Discovery, Paramount Pictures International, and Sony Pictures Releasing—as well as Roadshow Films, CMC Pictures, A24, VVS Films, and Icon Film Distribution Australia.

Additionally, the annual market share for 2025~2023 were used to estimate the proportion of the industry dominated by the Big Five.¹⁴⁷ These proportions, combined with studio stances, were used to develop scenario-based future projections on the state of the industry.

Table 6: Categorised Studios into the Four Stances

2025 Market Share Analysis (Screen Australia, 2025):¹⁴⁸

1. Walt Disney (24%)

2. Universal (20%)

3. Warner Bros. (19%)

4. Paramount (8%)

5. Sony (6%)

6. Roadshow (4%)

7. Other (19%)

77% of film market controlled by Big Five

2024 Market Share Analysis (Screen Australia, 2024):¹⁴⁹

1. Walt Disney (23%)

2. Universal (19%)

3. Warner Bros. (16%)

4. Sony (12%)

5. Paramount (9%)

6. Roadshow (4%)

7. Other (17%)

79% of film market controlled by Big Five

2023 Market Share Analysis (Screen Australia, 2023):¹⁵⁰

1. Universal (20%)

2. Warner Bros. (19%)

3. Walt Disney (19%)

4. Sony (12%)

5. Paramount (9%)

6. Roadshow (4%)

7. Other (17%)

79% of film market controlled by Big Five

The Big Five controlled approximately 77%-79% of annual market share between 2023 and 2025, with an average of 78%. As a result, the positions adopted by these major studios are highly determinative of the overall proportion of films covered by AI licensing agreements. 

Analysis reveals that the majority of major Hollywood studios are currently not engaging in publicly identifiable external AI licensing. Universal Pictures was categorised as ‘Restrictive,’ reflecting their restrictive approach towards unauthorised AI training through the use of “No AI Training” warnings in film credits starting June 2025, threatening to take legal action against unauthorised duplication or distribution of their work.¹⁵¹ Additionally, their litigation against MidJourney and MiniMax (joint lawsuit alongside Disney and Warner Bros.) for alleged copyright infringement may suggest the entity’s restrictive perception towards AI licensing.¹⁵²

Warner Bros. Discovery and Paramount Pictures were categorised under ‘Conditional Future,’ as their potential $111 billion merger in the third quarter of 2026 will create a massive combined intellectual property library, making it a potential source of AI training.¹⁵³ Following the merger, the consolidated entity may take a controlled, defensive response to protect their large IP library by engaging in AI licensing, and offer legal rights for AI-driven derivative creations for monetisation.

Walt Disney and Sony Pictures Releasing were overall classified as ‘Experimental/Controlled Adoption,’ given their internal use of AI technologies and their engagement with AI initiatives alongside litigation activity.¹⁵⁴ Although cancelled in March 2026, OpenAI and Disney announced that they were in the works of a partnership on December 11^th, 2025.¹⁵⁵ It would have been a landmark deal which would have granted OpenAI a three-year license to generate content using Disney film characters in its ‘Sora’ platform, a video generation tool.¹⁵⁶ Sony Music Publishing’s November 2025 licensing deal with Klay Vision Inc. in the music sector may imply positive attitudes towards external licensing in film in the future.¹⁵⁷ The remaining studios were placed in the ‘Limited Public Information’ category due to limited public information regarding external AI licensing of their film catalogues.

---

2: PROJECTION OF FUTURE LICENSING MARKET

A: MUSIC AND AUDIO

The state of the licensing markets appear to be somewhat mature, or at the very least, stagnant. Given that there have not been any reported new deals since January 2026, it is likely that the current state of licensing markets will maintain until June 2026. Sony’s lawsuits against Suno and Udio are still ongoing, and a fair-use ruling is expected in mid 2026, the outcome of which will undoubtedly carry huge implications for the future of AI music and licensing, while Universal may decide to continue its action against Udio later in 2026 in the absence of a settlement. In any event, barring any change to the market dominance of The Big 3, little is expected to change in terms of the overall percentage of works covered by AI licensing agreements between now and the end of 2026 or the end of 2028. It remains to be seen whether new, ‘ethically-trained’ AI music generation platforms unveiled recently or set to be released shortly will be able to successfully convince independent artists to ‘opt-in’.

B: WRITTEN WORKS

Whether the current market of direct licensing deals is likely to develop in the three years the government has purported not to intervene will involve many considerations. The success of current licensing deals, new deals entered into by major industry players, global responses to current copyright litigation and the attitude of AI developers to copyright laws are all likely to play a major role. 

For the news media industry, the illegal data scraping of the industry’s journalism is likely to hinder industry confidence when considering licensing deals. Additionally, whether the publishing industry will fairly compensate its authors for their works used to train AI models will be decisive in the success of these deals and the willingness of authors to opt-in to them. 

The nature of independent news media corporations and book publishers are likely to also play a significant role in the development of the licensing market, particularly as these smaller organisations do not have the ability to engage large technology companies to enter into licensing deals. Instead, relying on industry services will prove paramount to these organisations. In particular, the recent initiative by the UK based Publishers’ Licensing Services to establish an opt-in collective licensing scheme will prove paramount in the success of the evolving licensing market.¹⁵⁸

C: PHOTOGRAPHY AND FILM

Two primary scenarios were developed based on the observed stances of the examined studios, as well as their market share percentage, to project the future state of the film industry. 

SCENARIO A: LIMITED ADOPTION/CONSERVATIVE PROJECTION

Under this scenario, the Big Five studios refrain from entering into AI licensing agreements and instead resort to copyright enforcement through litigation. Other, more minor studios partially engage in AI licensing deals to monetise off of their film catalogue and generate supplementary revenue. Given that approximately 78% of the film industry is dominated by the Big Five, the proportion of theatrical films covered by AI licensing agreement is projected to remain low, around 5% to 10%. This scenario reflects the current phase of the film industry, characterised by litigation-driven resistance and cautious AI adoption. This realistically may represent the June 2026 state of the industry. 

SCENARIO B: CONDITIONAL EXPANSION

Under this hypothetical scenario, the Warner Bros. Discovery and Paramount Pictures merger is confirmed in late 2026. This merger would consolidate an expansive IP library, potentially creating incentives for the merged entity to monetise and protect their proprietary catalogue through AI licensing agreements. Meanwhile, experimental engagement from major studios such as Walt Disney may result in the partial adoption of licensing agreements, whereas entities such as Universal Pictures may continue to resist. The proportion of theatrical films covered by AI licensing agreements is projected to increase to approximately 35%~45% under this conditional scenario. This reflects the large impact that the Big Five and their actions will have on the film industry. 

SCENARIO C: GRADUAL INDUSTRY NORMALISATION (END OF 2028)

Continuing from Scenario B, by the end of the Productivity Commission’s three-year monitoring period, AI licensing may become a more established commercial practice within the film industry. Assuming that major ongoing copyright disputes have concluded and produced greater legal certainty regarding AI training, and that existing licensing agreements have demonstrated commercial value and remunerations for rightsholders—additional studios may be more willing to enter into agreements rather than primarily relying on litigation. However, absent policy intervention, legal uncertainties may continue to persist. As a result, AI licensing is unlikely to have developed into a universal industry standard and instead emerge as a commercial option for film studios. Under this scenario, approximately 45%~60% of films may be covered under AI licensing agreements by late 2028, reflecting gradual industry normalisation while still acknowledging that a significant proportion of films may remain unlicensed.

---

3: CONCLUSION AND RECOMMENDATION

This paper originally sought to determine what percentage of creative content is currently subject to licensing deals and the evolution of this percentage figure in the future. Based on limitations discussed throughout this paper, identified a significant issue with analysing the licensing deals between AI developers and creative industries. As researchers, we only have access to public information on licensing deals that are by their nature private commercial transactions. This should be taken into account by policymakers, particularly when responding to the PC Report.

While the Productivity Commission report recommends a three-year “wait and monitor” approach to AI and copyright licensing markets, it is unclear exactly how the Australian government intends to “monitor” these markets when such little detail is publicly available regarding licensing agreements between rights holders and AI developers. By what metric will a market-based solution be assessed to be ‘successful’ or ‘sufficient’? Are we to take the vaguely-worded yet optimistic-sounding press releases at face value? Without transparency, or at the very least, insight, it is unlikely that the situation can be “monitor[ed]” in any meaningful way.

Despite earlier claims by AI companies that licensing is ‘too difficult’ or will undermine their competitiveness, recent developments in licensing deals challenged this view. In the music sector, there has been a growing number of licensing agreements with major rightsholders, particularly over the past 6 months. This is likely to continue, as AI companies seek to avoid further costly lawsuits and fines. Although other creative sectors have yet to follow this pattern, our recommendations, as outlined below, set out the important considerations policy makers should take into account across the creative industries.

Recommendations:

1. Reject blanket TDM exceptions: market-led licensing framework should be favoured over broad copyright exemptions for AI training, as it preserves the value of creative works and avoid normalising unauthorised use. 

2. Establish pay-per-value compensation: support industry bodies in their lobbying efforts to establish pay-per-value compensation right, to ensure the original creators are compensated when AI outputs substantially rely on their specific work and style. 

3. Mandate standardised opt-in obligation: AI developers should implement transparent, standardised opt-in mechanisms for dataset provenance,¹⁵⁹ ensuring independent creators are not subsumed into buyout models without explicit consent. This approach would further allow regulators to effectively monitor the development of the commercial licensing market during the proposed three-year monitoring period.

4. Protect individual creators: the current licensing scheme covers primarily the major rightsholders, leaving the independent artists exposed to involuntary or unauthorised exploitation, which is evident in several ongoing class actions. A more ideal approach should address the structural imbalance embedded within the current voluntary licensing framework,¹⁶⁰ as well as developing opt-out mechanisms and data poisoning protections.

---

4: REFERENCES

1. Australian Productivity Commission, Harnessing data and digital technology (Inquiry Report No 111, December 2025) (‘PC Report’).

2. Ibid.

3. PC Report (n 1) 32.

4. Ibid.

5. Ibid.

6. Ibid 31.

7.  ‘AI Licensing by Copyright Owners’, Copyright Alliance (Web Page) <https://copyrightalliance.org/artificial-intelligence-copyright/licensing/copyright-owners/#music-audio-licensing>.

8.  Copyright Act 1968 (Cth), s10.

9. Ibid.

10. Ibid.

11.   Kristin Robinson, ‘Is Suno the Music Industry’s Biggest Nightmare — or Greatest Hope?’, Billboard (online, 11 March 2026) <https://www.billboard.com/pro/suno-ai-music-startup-cover-story/>.

12.  Hope King, ‘Exclusive: Gen AI music app Suno comes out of stealth’, Axios (online, 20 December 2023) <https://www.axios.com/2023/12/20/suno-gen-ai-music-microsoft>.

13.  Brian Hiatt, ‘AI-Music Arms Race: Meet Udio, the Other ChatGPT for Music’, Rolling Stone (online, 10 April 2024) <https://www.rollingstone.com/music/music-features/udio-ai-music-chatgpt-suno-1235001675/>.

14.  As per US law; Bruce Haring, ‘AI Startup Suno Claims “Fair Use” Copyright Doctrine Allows Training On Major Recordings’, Deadline (online, 1 August 2024) <https://deadline.com/2024/08/ai-startup-suno-claims-copyright-doctrine-allows-ai-training-1236029563/>.

15.  Recording Industry Association of America, ‘Record Companies Bring Landmark Cases for Responsible AI Against Suno and Udio in Boston and New York Federal Courts, Respectively’ (Press Release, 24 June 2024) <https://www.riaa.com/record-companies-bring-landmark-cases-for-responsible-ai-againstsuno-and-udio-in-boston-and-new-york-federal-courts-respectively/>.

16.  Deezer, ‘Deezer launches world’s first AI tagging system for music streaming’ (Media Release, 20 June 2025) <https://newsroom-deezer.com/2025/06/deezer-launches-worlds-first-ai-tagging-system-for-music-streaming/>; Spotify, ‘Spotify Strengthens AI Protections for Artists, Songwriters, and Producers’, (Media Release, 25 September 2025) <https://newsroom.spotify.com/2025-09-25/spotify-strengthens-ai-protections/>.

17.  Top Music Attorney, ‘I'm Suing Suno & Udio…’ (Video, 16 June 2025) <https://www.youtube.com/watch?v=VlKBpyjZU4I>.

18.  Ashley King, ‘Mutiny on Udio’s Deck: Subscribers Livid After Platform Disables Downloads Following UMG Deal’, Digital Music News (online, 31 October 2025) <https://www.digitalmusicnews.com/2025/10/31/udio-downloads-disabled-umg-deal/>.

19.  vAIsual, ‘vAIsual and Rightsify ink licensing deal to deliver royalty-free music for AI training’ (Media Release, 8 August 2023) <https://vaisual.com/vaisual-and-rightsify-ink-licensing-deal-to-deliver-royalty-free-music-for-ai-training/>.

20.  ‘Pasadena-based Rightsify Launches GCX, ‘World’s First AI Music Dataset Licensing Service’’, Pasadena Now (online, 24 August 2023) <https://pasadenanow.com/main/pasadena-based-rightsify-launches-gcx-worlds-first-ai-music-dataset-licensing-service>.

21.  Daniel Tencer, ‘Stability AI launches text-to-music generator trained on licensed content via a partnership with music library AudioSparx’, Music Business Worldwide (online, 14 September 2023) <https://www.musicbusinessworldwide.com/stability-ai-launches-text-to-music-generator-trained-on-licensed-content-via-a-partnership-with-music-library-audiosparx/>.

22.  Universal Music Group, ‘Universal Music Group and BandLab Technologies Announce First-of-Its-Kind Strategic AI Collaboration’ (Media Release, 18 October 2023) <https://www.universalmusic.com/universal-music-group-and-bandlab-technologies-announce-first-of-its-kind-strategic-ai-collaboration/>.

23.  Universal Music Group, ‘SoundLabs and Universal Music Group Announce Strategic Agreement to Offer Responsibly Trained AI Technology and Vocal Modeling Plug-In Micdrop to UMG Artists’ (Media Release, 18 June 2024) <https://www.universalmusic.com/soundlabs-and-universal-music-group-announce-strategic-agreement-to-offer-responsibly-trained-ai-technology-and-vocal-modeling-plug-in-micdrop-to-umg-artists/>.

24.  Record of the Day, ‘SourceAudio Teams Up with ElevenLabs as Preferred Partner for AI Music Licensing’ (Press Release, 23 July 2025) <https://www.recordoftheday.com/on-the-move/news-press/sourceaudio-teams-up-with-elevenlabs-as-preferred-partner-for-ai-music-licensing>.

25.  Murray Stassen, ‘Universal Music strikes strategic agreement with AI startup ProRata, which just raised $25m for a chatbot and tech to attribute and compensate content owners’, Music Business Worldwide (online, 6 August 2024) <https://www.musicbusinessworldwide.com/universal-music-strikes-strategic-agreement-with-ai-startup-prorata-which-just-raised-25m-for-a-chatbot-and-tech-to-attribute-and-compensate-content-owners1/>.

26.  Lauren Forristal, ‘Meta, Universal Music Group address AI music in new licensing agreement’ TechCrunch (online, 12 August 2024) <https://techcrunch.com/2024/08/12/meta-universal-music-group-address-ai-generated-content/>.

27.  Stuart Dredge, ‘Symphonic opens its catalogue up for licensed AI training’, Music Ally (online, 21 August 2024) <https://musically.com/2024/08/21/symphonic-opens-its-catalogue-up-for-licensed-ai-training/>.

28.  ElevenLabs, ‘Bertelsmann and ElevenLabs partner to support storytelling with AI’ (Media Release, 29 November 2024) <https://elevenlabs.io/blog/bertelsmann>.

29.  Tim Ingham, ‘Eleven Music, new AI rival to Suno, launches with Kobalt, Merlin deals – and warnings for users who want to infringe copyrights’, Music Business Worldwide (online, 5 August 2025) <https://www.musicbusinessworldwide.com/eleven-music-new-ai-rival-to-suno-launches-with-merlin-kobalt-licensing-deals-in-the-bag/>; Tim Ingham, ‘Exclusive: Kobalt’s AI deal with Suno rival ElevenLabs guarantees publishing ‘parity’ with recorded music’, Music Business Worldwide (online, 5 August 2025) <https://www.musicbusinessworldwide.com/kobalts-ai-suno-most-favored-nation/>.

30.  Mansoor Rahimat Khan, ‘Introducing maestro: Beatoven.ai’s New Foundation Model’, Beatoven.ai (online, 27 August 2025) <https://www.beatoven.ai/blog/introducing-maestro/>.

31.  STIM, ‘STIM Launches the World’s First AI License for Music’ (Media Release, 9 September 2025) <https://www.stim.se/en/news/stim-launches-the-worlds-first-ai-license-for-music>.

32.  'KDDI and Universal Music sign strategic partnership in music and entertainment', Telecompaper (online, 30 September 2025) <https://www.telecompaper.com/news/kddi-and-universal-music-sign-strategic-partnership-in-music-and-entertainment--1549582>.

33.  Daniel Tencer, ‘Major Music Publishers Ink AI Licensing Deals with Lyrics and Music Data Company Musixmatch’, Music Business Worldwide (online, 15 October 2025) <https://www.musicbusinessworldwide.com/major-music-publishers-ink-ai-licensing-deals-with-lyrics-and-music-data-company-musixmatch/>.

34.  Spotify, ‘Sony Music Group, Universal Music Group, Warner Music Group, Merlin, and Believe to Partner with Spotify to Develop Artist-First AI Music Products’ (online, 16 October 2025) <https://newsroom.spotify.com/2025-10-16/artist-first-ai-music-spotify-collaboration/>.

35.  Universal Music Group, ‘Universal Music Group and Udio Announce Udio’s First Strategic Agreements for New Licensed AI Music Creation Platform’ (online, 29 October 2025) <https://www.universalmusic.com/universal-music-group-and-udio-announce-udios-first-strategic-agreements-for-new-licensed-ai-music-creation-platform/>.

36.  Universal Music Group, ‘Universal Music Group and Stability AI Announce Strategic Alliance to Co-develop Professional AI Music Creation Tools’, PR Newswire (online, 30 October 2025) <https://www.prnewswire.com/news-releases/universal-music-group-and-stability-ai-announce-strategic-alliance-to-co-develop-professional-ai-music-creation-tools-302599789.html>.

37.  Universal Music Group (n 37).

38.  Warner Music Group, ‘Warner Music Group and Udio Collaborate to Build a New Licensed Music Creation Service’ (Press Release, 19 November 2025) <https://www.wmg.com/news/warner-music-group-and-udio-collaborate-to-build-a-new-licensed-music-creation-service>.

39.  Warner Music Group, ‘Warner Music Group and Stability AI Join Forces to Build the Next Generation of Responsible AI Tools for Music Creation’ (Press Release, 19 November 2025) <https://www.wmg.com/news/warner-music-group-and-stability-ai-join-forces-to-build-the-next-generation-of-responsible-ai-tools-for-music-creation>.

40.  Sony Music, ‘Music Technology Company Klay Signs First-Of-Its-Kind AI Licensing Deals with Universal Music Group, Universal Music Publishing Group, Sony Music Entertainment, Sony Music Publishing, Warner Music Group & Warner Chappell Music’ (online, 20 November 2025) <https://www.sonymusic.com/sonymusic/klay-signs-ai-licensing-deals-universal-sony-music-publishing-warner-chappell/>.

41.  Warner Music Group, ‘Warner Music Group and Forge Groundbreaking Partnership’ (Press Release, 25 November 2025) <https://www.wmg.com/news/warner-music-group-and-suno-forge-groundbreaking-partnership>.

42.  Universal Music Group, ‘Universal Music Group to Transform Music Experience for Billions of Fans with NVidia AI’ (online, 6 January 2026) <https://www.universalmusic.com/universal-music-group-to-transform-music-experience-for-billions-of-fans-with-nvidia-ai/>.

43.  Merlin, ‘Merlin and Udio Partner to Advance AI for Independent Music’ (online, 20 January 2026) <https://merlinnetwork.org/merlin-and-udio-partner-to-advance-ai-for-independent-music/>.

44.  Chris Eggertsen, ‘Record Label Market Share Year-End 2025: Taylor Swift Powers REPUBLIC to New High; Interscope & Atlantic Finish Strong’ (online, 28 January 2026). <https://radiox.cms.socastsrm.com/2026/01/28/record-label-market-share-year-end-2025-taylor-swift-powers-republic-to-new-high-interscope-atlantic-finish-strong/>.

45.  Sam C Shin, ‘The Role of Artificial Intelligence in Music Creation: Implications for Artists and the Industry’ (2024) 6(2) International Journal of Music Science, Technology and Art 38.

46.  Matthew Schurz, ‘Fair Use or Foul Play? Copyright Law’s Battle Over Using Sound Recordings in AI Training’ (2025) 48(1) UC Law SF Communication and Entertainment Journal 87.

47.  Copyright Act 1968 (Cth) s 31.

48.  Australian Productivity Commission, Harnessing data and digital technology (Inquiry Report No 111, December 2025) 34-35.

49.  News Corp, ‘News Corp and Open AI Sign Landmark Multi-Year Global Partnership’ (Media Release, 22 May 2024); OpenAI, ‘A landmark multi-year global partnership with News Corp’ (Media Release, 22 May 2024).

50.  News Corp, ‘News Corp and Open AI Sign Landmark Multi-Year Global Partnership’ (Media Release, 22 May 2024); Open AI, ‘A landmark multi-year global partnership with News Corp’ (Media Release, 22 May 2024).

51.  Shirin Ghaffray, ‘News Corp signs lucrative content deal with ChatGPT maker OpenAI’, The Sydney Morning Herald (online, 23 May 2024) < https://www.smh.com.au/business/companies/news-corp-inks-deal-with-chatgpt-maker-openai-20240523-p5jfwq.html>.

52.  Alexandra Burell, ‘News Corp, Meta in AI Content Licensing Deal Worth Up to $50 Million a Year’, The Wall Street Journal (online, 3 March 2026) < https://www.wsj.com/business/media/news-corp-meta-in-ai-content-licensing-deal-worth-up-to-50-million-a-year-d4fbf244>.

53.  The Guardian, ‘Guardian Media Group announces strategic partnership with OpenAI (Medial Release, 15 February 2025); OpenAI, ‘OpenAI and Guardian Media Group launch content partnership’ (Media Release, 14 February 2025).

54.  ‘Scaling AI for everyone’, OpenAI (online, 27 February 2026) < https://openai.com/index/scaling-ai-for-everyone/>.

55.  Jonathan Barrett, ‘Nine urges Albanese to force tech companies to compensate media in face of AI threat’, Guardian Australia (online, 24 February 2026) < https://www.theguardian.com/media/2026/feb/24/nine-entertainment-australia-news-bargaining-incentive-ai-threat>.

56.  Zoe Samios, ‘”He’s got our back”: Nine CEO says PM will resolve tech giant payments’, Australian Financial Review (online, 24 February 2026) < https://www.afr.com/companies/media-and-marketing/nine-profit-climbs-as-english-premier-league-drives-subscriber-growth-20260224-p5o4vu>.

57.  Elias Visontay and Brittany Busch, ‘Nine accuses AI firms of ‘illegal scraping’ of journalism’, The Sydney Morning Herald (online, 27 October, 2025) < https://www.smh.com.au/politics/federal/very-clear-labor-rules-out-exemption-for-ai-giants-to-take-content-for-free-20251027-p5n5hs.html>.

58.  Sam Buckingham-Jones, ‘AI firms crawling Nine Entertainment’s news sites 10 times a second’, Australian Financial Review (online, 6 August, 2025) < https://www.afr.com/companies/media-and-marketing/ai-firms-crawling-nine-entertainment-s-news-sites-10-times-a-second-20250806-p5mkop>.

59.  Jonathan Barrett, ‘Nine urges Albanese to force tech companies to compensate media in face of AI threat’, Guardian Australia (online, 24 February 2026) < https://www.theguardian.com/media/2026/feb/24/nine-entertainment-australia-news-bargaining-incentive-ai-threat>; Zoe Samios, ‘”He’s got our back”: Nine CEO says PM will resolve tech giant payments’, Australian Financial Review (online, 24 February 2026) < https://www.afr.com/companies/media-and-marketing/nine-profit-climbs-as-english-premier-league-drives-subscriber-growth-20260224-p5o4vu>.

60.  John Buckley, ‘Seven, ABC, SBS and ACM form alliance to chase AI deals’, Capital Brief (online, 28 November 2024) < https://www.capitalbrief.com/article/seven-abc-sbs-and-acm-form-alliance-to-chase-ai-deals-4df1265e-8380-4238-8c50-6ff708b0d892/>

61.  Sam Buckingham-Jones, ‘Don’t build trillion-dollar AI companies using our content: ABC boss’, Australian Financial Review (online, 8 August 2025) < https://www.afr.com/companies/media-and-marketing/don-t-build-trillion-dollar-ai-companies-using-our-content-abc-boss-20250807-p5ml9o>.

62.  Ariel Bogle, ‘New York Times, CNN and Australia’s ABC block OpenAI’s GPTBot web crawler from accessing content’, Guardian Australia (online, 25 August 2023) < https://www.theguardian.com/technology/2023/aug/25/new-york-times-cnn-and-abc-block-openais-gptbot-web-crawler-from-scraping-content>.

63.  ‘HarperCollins AI deal: what authors need to know’, Australian Society of Authors (online, 21 November 2024) < https://www.asauthors.org.au/news/harpercollins-ai-deal-what-authors-need-to-know/>.

64.  Ella Creamer, ‘HarperCollins to allow tech firms to use its books to train AI models’, Guardian Australia (online, 20 November 2024) < https://www.theguardian.com/books/2024/nov/19/harpercollins-tech-firms-books-train-ai-models-nonfiction-artificial-intelligence>.

65.  ‘HarperCollins AI deal: what authors need to know’, Australian Society of Authors (online, 21 November 2024) < https://www.asauthors.org.au/news/harpercollins-ai-deal-what-authors-need-to-know/>.

66.  Ella Creamer, ‘HarperCollins to allow tech firms to use its books to train AI models’, Guardian Australia (online, 20 November 2024) < https://www.theguardian.com/books/2024/nov/19/harpercollins-tech-firms-books-train-ai-models-nonfiction-artificial-intelligence>.

67.  ‘HarperCollins AI deal: what authors need to know’, Australian Society of Authors (online, 21 November 2024) < https://www.asauthors.org.au/news/harpercollins-ai-deal-what-authors-need-to-know/>.

68.  ‘Penguin Random House UK Response to Government Consultation on Copyright and AI, Penguin Random House UK, (online, 25 February 2025) < https://www.penguin.co.uk/about/company-articles/response-to-consultation-copyright-and-ai>.

69.  Penguin Random House US, ‘PRH Joins Worldwide Coalition Condemning Unlicensed Use of Creative Works for Training Generative AI’ (Media Release, November 2024).

70.  Matilda Battersby, ‘Penguin Random House underscores copyright protection in AI rebuff’, The Bookseller (online, 18 October 2024) <https://www.thebookseller.com/news/penguin-random-house-underscores-copyright-protection-in-ai-rebuff>.

71.  ‘AI Position’, Hachette Australia (online) < https://www.hachette.com.au/ai-position/>.

72.  Blake Brittain, ‘Major publishers sue Meta for copyright infringement over AI training’, Reuters (online, 6 March 2026) < https://www.reuters.com/sustainability/boards-policy-regulation/major-publishers-sue-meta-copyright-infringement-over-ai-training-2026-05-05/>.

73.  Association of American Publishers (AAP), ‘Publishers Respond to Google’s Opposition in Class Action Infringement Suit’ (Media Release, 6 February 2026).

74.  ‘Our approach to AI at Pan Macmillan’, Pan Macmillan (online) <https://www.panmacmillan.com/ai-at-pan-macmillan>.

75.  Blake Brittain, ‘Major publishers sue Meta for copyright infringement over AI training’, Reuters (online, 6 March 2026) < https://www.reuters.com/sustainability/boards-policy-regulation/major-publishers-sue-meta-copyright-infringement-over-ai-training-2026-05-05/>.

76.  ‘The ‘Make it Fair’ campaign: a powerful call to safeguard creative rights in AI’, Pan Macmillan (online, 25 February 2025) < https://www.panmacmillan.com/news/the-make-it-fair-campaign>.

77.  ‘Pan Macmillan welcomes government decision to move away from copyright exception for AI training’, Pan Macmillan (online, 18 March 2026) <https://www.panmacmillan.com/news/pan-macmillan-welcomes-government-decision-to-move-away-from-copyright-exception-for-ai-training>.

78.  Cade Metz et al, ‘How Tech Giants Cut Corners to Harvest Data for AI’, The New York Times (online, 6 April 2024) < https://www.nytimes.com/2024/04/06/technology/tech-giants-harvest-data-artificial-intelligence.html>.

79.  Ella Creamer, ‘Meta ‘discussed buying publisher Simon & Schuster to train AI’, Guardian Australia (online, 10 April 2024) <https://www.theguardian.com/books/2024/apr/09/meta-discussed-buying-publisher-simon-schuster-to-train-ai>.

80.  ‘An update on the latest AI developments’, Australian Society of Authors (online, 17 April 2024) <https://www.asauthors.org.au/news/an-update-on-the-latest-ai-developments/>.

81.  Dilan Thampapillai, ‘Black Inc has asked authors to sign AI agreement. But why should writers help AI learn how to do their job?’, The Conversation (online, 6 March 2025) < https://theconversation.com/black-inc-has-asked-authors-to-sign-ai-agreements-but-why-should-writers-help-ai-learn-how-to-do-their-job-251467>.

82.  ‘Black Inc. AI licensing – what authors need to know’, Australian Society of Authors (online, 5 March 2025) < https://www.asauthors.org.au/news/black-inc-ai-licensing-what-authors-need-to-know/>.

83.  Ibid; Kelly Burke, ‘’Sign our own death warrant’: Australian writers angry after Melbourne publisher asks them to sign AI agreements’, Guardian Australia (online, 4 March 2025) < https://www.theguardian.com/books/2025/mar/05/black-inc-melbourne-publisher-ai-agreements-writers-anger>.

84.  Jonathan Barrett, ‘Nine urges Albanese to force tech companies to compensate media in face of AI threat’, Guardian Australia (online, 24 February 2026) < https://www.theguardian.com/media/2026/feb/24/nine-entertainment-australia-news-bargaining-incentive-ai-threat>; Zoe Samios, ‘”He’s got our back”: Nine CEO says PM will resolve tech giant payments’, Australian Financial Review (online, 24 February 2026) < https://www.afr.com/companies/media-and-marketing/nine-profit-climbs-as-english-premier-league-drives-subscriber-growth-20260224-p5o4vu>.

85.  AAP, ‘AAP Will Provide Its Fact-based News Content to Google’ (online, 18 August 2025) <https://www.aap.com.au/about/announcements/aap-partners-with-google-to-bring-trusted-news-to-gemini-app/>.

86.  Canva, ‘Expanding Our Partnership with Getty Images to Strengthen Our AI’ (online, 27 July 2024) <https://www.canva.com/newsroom/news/getty-ai-partnership/>; Shutterstock, ‘Shutterstock Partners with OpenAI and Leads The Way to Bring AI-generated Content to All’ (online, 25 October 2022) <https://investor.shutterstock.com/node/12506/pdf>; Shutterstock, ‘Shutterstock Expands Partnership with OpenAI, Signs New Six-Year Agreement to Provide High-Quality Training Data’ (online, 11 July 2023) <https://investor.shutterstock.com/news-releases/news-release-details/shutterstock-expands-partnership-openai-signs-new-six-year>; Bridget Johnston et al, ‘LG AI Research Develops Differentiated AI Technologies and Partners with Shutterstock to Make The World More Productive and Creative’ Shutterstock (online, 24 August 2023) <https://www.shutterstock.com/blog/lg-partners-with-shutterstock-to-advance-ai-for-better-life?dd_referrer=https%3A%2F%2Fwww.google.com%2F>; Shutterstock, ‘Shutterstock Expands Long-standing Relationship with Meta’ (online, 12 January 2023) <https://investor.shutterstock.com/news-releases/news-release-details/shutterstock-expands-long-standing-relationship-meta>; AAP, ‘AAP Will Provide Its Fact-based News Content to Google’ (online, 18 August) <https://www.aap.com.au/about/announcements/aap-partners-with-google-to-bring-trusted-news-to-gemini-app/>; Getty Images, ‘Getty Images and Perplexity Strike Multi-year Image Partnership’ (online, 31 October 2025) <https://investors.gettyimages.com/news-releases/news-release-details/getty-images-and-perplexity-strike-multi-year-image-partnership>.

87.  Halvor Manshaus, ‘AI In the Training Room – Getty Images vs. Stability AI Limited’ (online, 9 December 2025) Schjodt < https://schjodt.com/news/ai-in-the-training-room-getty-images-vs-stability-ai-limited#:~:text=by%20Halvor%20Manshaus,Bakground%20to%20the%20parties>.

88.  Ibid; John Fairbairn et al, ‘AI Copyright Litigation in Australia’ (online, 6 November 2025) <https://www.minterellison.com/articles/ai-copyright-litigation-in-australia>.

89.  Zach Schor, ‘Andersen v. Stability AI: The Landmark Case Unpacking the Copyright Risks of AI Images Generators’ Journal of Intellectual Property & Entertainment Law (online, 2 December 2024) https://jipel.law.nyu.edu/andersen-v-stability-ai-the-landmark-case-unpacking-the-copyright-risks-of-ai-image-generators/; Mary Bennett et al, ‘Ongoing Legal Battles Over AI Copyright Infringement: Implications for Tech Companies’ (online, 28 August 2024) <https://www.jdsupra.com/legalnews/ongoing-legal-battles-over-ai-copyright-2160521/#:~:text=Stability%20AI%2C%20alongside%20others%2C%20relies%20on%20extensive,under%20legal%20scrutiny%20for%20alleged%20copyright%20violations.>.

90.  Hagens Berman, ‘CapCut User Rights’ (online, 1 August 2023) <https://www.hbsslaw.com/cases/capcut-user-rights>.

91.  Tracy Bagdonas, ‘Class Action Lawsuit: Content Creators Claims Snap Wrongfully ‘Scrapes’ Content From YouTube to Train AI’ (online, 2 February 2026) <https://www.classaction.org/news/class-action-lawsuit-content-creators-claim-snap-wrongfully-scrapes-content-from-youtube-to-train-ai>.

92.  Blake Brittain, ‘Meta, Nvidia, Roblox sued by 3D Artist Over AI Training’ Reuters (online, 26 March 2026) <https://www.reuters.com/legal/government/meta-nvidia-roblox-sued-by-3d-artist-over-ai-training-2026-03-26/>.

93.  Penny Williams, ‘The impact of disruptive innovation on creative workers: the case of photographers’ (2021) 14(2) Creative Industries Journal 130–151.

94.  Paula McDonald et al, ‘How professional photographers engage with and resist digital platform work’ in (2021) 23(6) New Media and Society 1602–1623.

95.  Image Makers Association, ‘Online Member Questionnaire: Australian Image Makers & Their Income’ (online) <https://www.imagemakersassociation.com.au/chargingsurvey>.

96.  Getty Images, ‘Website Terms’, <https://www.gettyimages.ae/company/terms>.

97. Ibid.

98. Ibid.

99. Getty Images, ‘Royalties Guide’, < https://contributors.gettyimages.com/help/article/5191>.

100. Getty Images, ‘Generate new AI images or modify our creative imagery’, <https://www.gettyimages.co.uk/ai>.

101. Canva, ‘Terms of Use’ <https://www.canva.com/policies/terms-of-use/>.

102. Ibid.

103. Ibid.

104. Adobe, ‘Content Credentials Generative AI Training and Usage Preference’ (online, 2 September 2025) <https://helpx.adobe.com/creative-cloud/apps/adobe-content-authenticity/generative-ai-training-preferences.html>.

105. Adobe, ‘Solution Brief: Generative AI Built for Business’ (online) <https://www.adobe.com/content/dam/cc/en/trust-center/ungated/whitepapers/corporate/adobe-gen-ai-built-for-business-solution-brief.pdf#:~:text=For%20these%20models%2C%20we%20create%20image%20datasets,copyright%20has%20expired.%20Why%20does%20this%20matter?>.

106. Hagens Berman, ‘CapCut User Rights’ (online, 1 August 2023) <https://www.hbsslaw.com/cases/capcut-user-rights>.

107. Pear Collective Photography, ‘The Impact of Social Media on Photographers’ (online, 30 November 2025) <https://www.pearlcollectivephotography.com/blog/the-impact-of-social-media-on-photographers>.

108. Joe tidy, ‘How Your Personal Data is Being Scraped From Social Media’ BBC News (online, 16 July 2021) <https://www.bbc.com/news/business-57841239>.

109. Meltwater, Digital 2026: Global Overview Report (Report) 384.

110. Content Hype, ‘Short-Form Video Marketing in Australia: Stats, Trends, and a 1-Hour Production Hack’ (Online) <https://contenthype.com.au/2025/06/08/short-form-video-marketing-in-australia-stats-trends-and-a-1-hour-production-hack/>.

111.  Ibid.

112.  Meta, ‘Terms of Service Overview’ (online, 1 January 2025) <https://www.facebook.com/terms/> clause 3.3 (2).

113. Ibid.

114. Instagram, ‘Terms of Use’ (online) <https://help.instagram.com/478745558852511>; Alexandra Popova, ‘Ten Things You Should Know About Instagram’s Terms of Use’ (Online, 12 September 2018) <https://theconversation.com/ten-things-you-should-know-about-instagrams-terms-of-use-102800#:~:text=Instagram%20claims%20it%20does%20not,Instagram%20would%20breach%20that%20licence.>; Ruchika, ‘Who Owns the Content You Post on Instagram’ (online, 29 June 2025) <https://ai.lawctopus.com/who-owns-the-content-you-post-on-instagram/>.

115.  Meta, ‘How Meta Uses Information for Generative AI Models and Features’ (online) <https://www.facebook.com/privacy/genai/>.

116.  Mike Clark, ‘Privacy Matters: Meta’s Generative AI Features’ in Meta (online, 27 September 2023) <https://about.fb.com/news/2023/09/privacy-matters-metas-generative-ai-features/>.

117.  Copyright Agency, Submission to CAIRG, Response to CAIRG Copyright and AI Consultation Paper (December 2025) < https://static-copyright-com-au.s3.amazonaws.com/uploads/2026/02/R02846-CAIRG-AI-consultation-paper-response-2025-12-11.pdf>.

118.  Copyright Agency, Submission to Joint Select Committee,  Submission to Joint Select Committee on Social Media and Australian Society (July 2024) < https://static-copyright-com-au.s3.amazonaws.com/uploads/2024/07/R02742-Parl-inquiry-into-Social-Media.pdf>.

119.  Coherent Market Insights, ‘Stock Photography Market to Reach US$4.8 Billion by 2028 – Coherent Market Insights’ (online, 30 July 2021) <https://www.coherentmarketinsights.com/press-release/stock-photography-market-3797#:~:text=The%20global%20stock%20photography%20market,%2C%20marketing%2C%20and%20customer%20acquisitions.>; Custom Market Insight, ‘Global Stock Images Market 2024–2033’ (online) <https://www.custommarketinsights.com/report/stock-images-market/>.

120.  David Kennedy, ‘REPORT: Productivity Commission Misfires on Content Ownership in the AI Economy’ Venture Insights (online, 11 August 2025) <https://www.ventureinsights.com.au/product/report-productivity-commission-misfires-on-content-ownership-in-the-ai-economy/?pdf=get-download> 8–9; Getty Images, Submission to the ACCC Digital Platforms Inquiry Issues Paper (10 April 2018) <https://www.accc.gov.au/system/files/Getty%20Images%20%28April%202018%29.pdf#:~:text=Getty%20Images%20serves%20nearly%201%20million%20customers,50%2C000%20in%20sport%20and%2040%2C000%20in%20news.>.

121.  Eliza Berlage, ‘Artists Work at Risk as Researcher Urges Privacy Checks over AI Caricature Trend’ ABC News (online, 3 March 2026) <https://www.abc.net.au/news/2026-03-03/ai-art-caricature-impact-on-creative-workers/106382724>.

122.  Copyright Agency (n 120).

123. Runway, ‘Runway Partners with Lionsgate’ (online, 18 September 2024) <https://runwayml.com/news/runway-partners-with-lionsgate>. 

124.  Mackenzie Shane, 'When AI Draws Batman: Are Generative AI Platforms Liable for Copyright Infringement?', Brooklyn Sports & Entertainment Law Blog (Web Page, 5 February 2026) <https://sports-entertainment.brooklaw.edu/film-tv/when-ai-draws-batman-are-generative-ai-platforms-liable-for-copyright-infringement/>.

125.  Disney Enterprises, Inc. et al v. Midjourney, Inc., No. 2:25-cv-05275-JAK-AJR (C.D. Cal. 2025) (Universal City Studios Productions LLLP, Marvel Characters Inc, Lucasfilm Ltd, Twentieth Century Fox Film Corporation and DreamWorks Animation as co-plaintiffs) (filed 11 June 2025); Warner Bros. Entertainment Inc. v. Midjourney, Inc., No. 2:25-cv-08376 (C.D. Cal. 2025) (filed 4 September 2025).

126.  Ibid. 

127.  Michael Nurse et al, ‘When AI Meets IP: The Disney and Universal v Midjourney Case’, Colin Biggers & Paisley (online, 28 August 2025) <https://www.cbp.com.au/insights/publications/when-ai-meets-ip-the-disney-and-universal-v-midjourney-case>.

128.  Jessica Riga, ‘Disney and Universal Sue AI Firm Midjourney for Copyright Infrigement’, ABC (online, 12 June 2025) <https://www.abc.net.au/news/2025-06-12/disney-and-universal-sue-ai-firm-midjourney-for-copyright/105407174>.

129.  17 USC § 107 (1976).

130.  Disney Enterprises, Inc. v. Minimax, No. 2:25-cv-08768 (C.D. Cal. 2025) (also naming Universal City Studios Productions LLLP, Warner Bros. Entertainment Inc, Marvel Characters Inc, Lucasfilm Ltd, Twentieth Century Fox Film Corporation, DC Comics, The Cartoon Network Inc, Turner Entertainment Co, Hanna-Barbera Productions Inc, and DreamWorks Animation as co-plaintiffs) (filed 16 September 2025).

131.  Rahul Vadisetty, 'Navigating Copyright and Fair Use in AI Training Data: Legal Challenges and Future Solutions' (Conference Paper, 2025 International Conference on Advances in Smart Systems, Innovations and Computing (ASSIC), Institute of Electrical and Electronics Engineers, 2025) <https://doi.org/10.1109/ASSIC64892.2025.11158012>. 

132.  Robert H Aft, From Script to Screen: The Importance of Copyright in the Distribution of Films (World Intellectual Property Organization, August 2011) <https://doi.org/10.34667/tind.28652>. 

133. Ibid. 

134.  Eleonora Rosati, 'The Future of the Movie Industry in the Wake of Generative AI: A Perspective under EU and UK Copyright Law' (2025) 59 Computer Law & Security Review 106207 <https://doi.org/10.1016/j.clsr.2025.106207>.

135.  Art P. ‘What Universal Pictures is Really Worried About with AI’ (online, 9 August 2025) <https://www.linkedin.com/pulse/what-universal-pictures-really-worried-ai-art-proctor-yrbzc?utm_source=share&utm_medium=guest_desktop&utm_campaign=copy>.

136.  Benjamin Dodman et al, ‘Cannes 2025: Cinema Urged to Ride ‘Unstoppable’ AI Wave As Critics Warn of Slippery Slope’, Finance 24 (online, 19 May 2025) <https://www.france24.com/en/culture/20250519-cannes-2025-film-industry-urged-ride-unstoppable-artificial-intelligence-critics-warn-slippery-slope-ai>.

137.  Paul Lehrman, et al., v. Lovo, Inc. (D SD, 1:24-CV-3770 (JPO)). 

138.  Steven Zeitchik, ‘The AI Movie Factory is Ramping Up’, The Hollywood Reporter (online, 10 September 2025) <https://www.hollywoodreporter.com/business/digital/ai-movie-factory-is-ramping-up-1236366567/>.

139.  Screen Actors Guild–American Federation of Television and Radio Artists, ‘SAG-AFTRA Members Approve 2023 TV/Theatrical Contracts Tentative Agreement’ (5 December 2023) <https://www.sagaftra.org/sag-aftra-members-approve-2023-tvtheatrical-contracts-tentative-agreement>.

140.  AIMICI, ‘AI Legal Cases 2025: What Film Producers Need to Know’ (June 2025) <https://aimici.co.uk/ai-legal-shake-up-what-recent-court-cases-mean-for-your-next-production/>.

141.  Jamie Vickers et al, ‘What AI Could Mean for Film and TV Production and the Industry’s Future’, Mckinsey & Company (Report, 23 January 2026) <https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/what-ai-could-mean-for-film-and-tv-production-and-the-industrys-future>.

142.  Media Entertainment & Arts Alliance, ‘SAG-AFTRA Strike Q&A’ (Web Page, 10 August 2023) <https://www.meaa.org/news/sag-aftra-strike-qa/>.

143. Kevin Madigan, 'Mid-Year Review: AI Copyright Case Developments in 2025', Copyright Alliance (Web Page, 21 August 2025) <https://copyrightalliance.org/ai-copyright-case-developments-2025/>. 

144.  FTI Consulting, 'FTI Consulting Reveals Hurdles for Video Publishers' (Web Page, 9 October 2024) <https://www.fticonsulting.com/insights/articles/fti-consulting-reveals-hurdles-video-publishers>.

145.  Mackenzie Shane, 'When AI Draws Batman: Are Generative AI Platforms Liable for Copyright Infringement?', Brooklyn Sports & Entertainment Law Blog (Web Page, 5 February 2026) <https://sports-entertainment.brooklaw.edu/film-tv/when-ai-draws-batman-are-generative-ai-platforms-liable-for-copyright-infringement/>.

146. 'Australian Box Office for 2025', Box Office Mojo (Web Page, 2025) <https://www.boxofficemojo.com/year/2025/?area=AU&sort=rank&ref_=bo_yld__resort>.

147.  Screen Australia, 'Cinema Industry Trends: Distributor Box Office Share' (Web Page, 2025) <https://www.screenaustralia.gov.au/insights-and-trends/cinema-industry-trends/australian-box-office/>.

148. Ibid.

149. Ibid.

150. Ibid.

151.  Winston Cho, 'Universal Pictures to Big Tech: We’ll Sue if You Steal Our Movies for AI', The Hollywood Reporter (Web Page, 6 August 2025) <https://www.hollywoodreporter.com/business/business-news/universal-pictures-big-tech-well-sue-if-you-steal-movies-ai-1236337712/>.

152.  'Disney, Universal, Warner Bros Discovery Sue China’s MiniMax for Copyright Infringement', Reuters (Web Page, 17 September 2025) <https://www.reuters.com/legal/litigation/disney-universal-warner-bros-discovery-sue-chinas-minimax-copyright-infringement-2025-09-16/>.

153. Rory Morris and Nardine Saad, 'Hollywood Stars Unite to Oppose Paramount and Warner Bros Discovery Merger', BBC (Web Page, 14 April 2026) <https://www.bbc.com/news/articles/cz0egvr94zro>.

154.  Noor Al-Sibai, 'Disney’s Secret Experiments with AI Have Reportedly Been a Comical Disaster', Futurism (Web Page, 6 August 2025) <https://futurism.com/disney-secret-ai-disaster>.

155.  OpenAI, 'The Walt Disney Company and OpenAI Reach Landmark Agreement to Bring Beloved Characters from Across Disney’s Brands to Sora' (Web Page, 11 December 2025) <https://openai.com/index/disney-sora-agreement/>.

156.  Gabrielle Pesantez, 'Authorized Imagination: Disney’s Partnership with OpenAI', The Columbia Journal of Law & the Arts (Web Page, 16 February 2026) <https://journals.library.columbia.edu/index.php/lawandarts/announcement/view/861>.

157.  Sony Music, 'Music Technology Company Klay Signs First-of-its-Kind AI Licensing Deals with Universal Music Group, Universal Music Publishing Group, Sony Music Entertainment, Sony Music Publishing, Warner Music Group & Warner Chappell Music' (Web Page, 20 November 2025) <https://www.sonymusic.com/sonymusic/klay-signs-ai-licensing-deals-universal-sony-music-publishing-warner-chappell/>.

158.  Lauren Brown, ‘PLS launches collective AI licensing scheme for publishers’, The Bookseller (online, 10 March 2026) < https://www.thebookseller.com/news/pls-launches-collective-ai-licensing-scheme-for-publishers>; Publishers’ Licencing Services, ‘PLS offers new collective AI licensing opportunity to publishers’ (Media Release, March 2026).

159.  See, e.g., Jerry Heinz, ‘Why Provenance Is the Foundation of Studio-Grade AI Production’ (online, 20 March 2026) <https://genvid.com/2026/03/20/why-provenance-is-the-foundation-of-studio-grade-ai-production/>: the integration of provenance tracking systems such as C2PA (Content Credentials) and IPTC Digital Source Type could improve transparency by providing clear identification of creative works and how they are utilised in AI training, replicating the ‘Chain of Title’ method of documentation within the film industry.

160.  Penelope Benton, National Association of the Visual Arts, ‘NAVA Welcomes Government Decision to Reject AI Copyright Exemption’ (Media Release, 27 October 2025) <https://visualarts.net.au/news-opinion/2025/nava-welcomes-government-decision-protect-artists-rights-ai-copyright-ruling/>.