Proposed Text and Data Mining Exception in Australia for AI systems: Implications

The Australian Productivity Commission’s interim report proposes a new “fair dealing” exception to allow text and data mining (TDM) for AI training without the explicit permission or compensation of copyright holders. This article examines the rationale behind the proposal, prevailing legal frameworks, and the widespread opposition it has triggered across Australia’s creative industries. Through a policy lens, it critiques both the economic arguments and the cultural risks, highlighting the urgent need for a balanced copyright framework that supports innovation while protecting creators’ livelihoods.

1. Introduction

The Productivity Commission’s Harnessing Data and Digital Technology interim report explores how artificial intelligence (AI) can drive economic productivity in Australia—estimating a boost of A$116 billion over the next decade. Among its various regulatory considerations, the Commission proposes introducing a TDM exception to the Australian Copyright Act, potentially permitting AI developers to mine copyrighted content under fair dealing, provided they use legally obtained and paid-for copies.This proposal, however, has drawn fierce criticism from authors, publishers, musicians, industry bodies, and advocates, who warn that it undermines creators’ rights and risks Australia’s cultural economy.

2. Current Legal Framework and Proposal Context

Under the Copyright Act 1968 (Cth), there is currently no legal exception for text or data mining. Reproducing or converting copyrighted materials without permission may constitute infringement. Existing fair dealing exceptions—such as for “research or study”—are unlikely to cover TDM, given that such processes often require copying entire works, exceeding the “reasonable portion” limit.

The Productivity Commission argues that AI developers are already training models on copyrighted materials—often without consent or compensation—and that this legal uncertainty hampers both research and commercial AI development in Australia. Its suggested solution? Expanding fair dealing to include TDM, so long as AI entities use legitimately acquired copies and could, in theory, be subject to licensing or collections-based arrangements akin to music licensing bodies.

3. Economic Rationale: Innovation vs. Creator Rights

Proponents of the TDM exception emphasize the economic promise of AI innovation, with projections of $116 billion in productivity gains—including wage increases of roughly A$4,300 per person over the next decade. Figures like Atlassian co-founder Scott Farquhar have argued that such copyright reform could attract billions of dollars in foreign investment.However, critics argue this economic case is both broad and speculative. The Association of Artist Managers (AAM) points out that the Commission’s A$116 billion figure reflects aggregate AI-driven gain—including gains from efficiency, automation, and new business models—not exclusively from TDM or copyright reforms—and lacks a detailed cost-benefit analysis specific to creative content.

4. Creative Industries Pushback

Across the cultural sector, there is overwhelming rejection of the proposal:

•	Artists and Rights Organisations: Arts Law published strong dismay, warning that weakening copyright would harm artists, especially First Nations creators—citing cultural misappropriation and dilution of Indigenous Cultural and Intellectual Property (ICIP).•	Authors: The Australian Society of Authors (ASA) and figures like Lucy Hayward label the proposal a “free pass” to Big Tech and “wage theft,” particularly problematic given that Australian writers earn an average of just ~$18,500 annually.•	Publishers: The Australian Publishers Association stresses that the exception would undermine the very frameworks that support creators’ livelihoods and cultural sustainability.•	Music Industry: ARIA and PPCA warn that unrestricted access would devastate the value of Australian cultural output—pointing to robust existing licensing systems that already facilitate AI use with permission.•	Wider Creative Sector: The Australian Writers’ Guild, BookPeople, APRA AMCOS, and others have characterized the proposal as institutionalizing theft and decimating the foundations of creative industries.•	Media & Journalistic Bodies: International organizations—including the European Publishers Council and News/Media Alliance—warn the proposal threatens journalism and democratic accountability, likening it to legitimizing digital piracy.•	Public Sentiment & Industry Leaders: Polls suggest 74% of Australians oppose AI companies using original work without consent or payment, and business chiefs like Commonwealth Bank’s CEO have flagged these exceptions as unfair and ethically risky—even amid AI adoption.
5. Legal and Cultural Considertions
Legal Complexity
Although TDM is often framed as a non-expressive, transformative use, Australian law currently lacks clarity on whether such use can be treated as fair dealing. The Australian Law Reform Commission notes that while some jurisdictions are debating TDM exceptions, Australia currently offers no specific provision, leaving legal uncertainty unresolved.
Cultural Risks and Equity
Critics warn that a TDM exception would disproportionately benefit multinational tech firms that scrape and monetize creative content, while Australian creators—who generate cultural value and economic benefit—receive nothing. This dynamic could exacerbate global inequities in intellectual property value and creative agency.
Moreover, for sectors like Indigenous arts, the implications extend beyond economics—potentially misappropriating culturally sensitive content absent proper consent or contextual oversight.
6. Policy Considerations and Alternatives
A balanced path forward could include:
•	Enhanced Licensing Mechanisms: Instead of sweeping exceptions, the Commission might prioritize improving and modernizing licensing frameworks—ensuring tech firms can access content with proper remuneration and consent, safeguarding creators’ rights.•	Collections-Based Approach: Drawing on models such as music collecting societies, a TDM licensing body could aggregate permissions and distribute royalties—harnessing efficiency without undercutting creators.•	Targeted Exceptions for Non-Creative Domains: If reform is necessary for certain contexts (e.g., scientific or medical research), exceptions could be narrowly tailored—excluding creative industries and sensitive cultural content.•	Clear Regulatory Guidelines: Improved clarity on definitions of “fair dealing,” “transformative use,” and consent mechanisms could reduce uncertainty while upholding ownership rights.
Such approaches would support innovation and AI development without sacrificing creative livelihoods or market balance.
Comment
The Productivity Commission’s proposed TDM exception is a flashpoint in Australia’s policy debates on AI, culture, and copyright. On one side, there's the promise of economic growth through AI productivity. On the other, there is a unified and principled backlash from Australia’s creative industries—warning of cultural erosion and economic inequity.At its core, this debate may well define whether Australia can leverage AI responsibly—ensuring technological advancement doesn't come at the cost of eroding artistic labor, cultural sovereignty, and creative diversity. Thoughtful policy design—focused on licensing frameworks, targeted exceptions, and fair compensation—offers a path toward innovation that respects the rights and dignity of creators.
Proposed Text and Data Mining Exception in Australia for AI systems: Implications of the Productivity Commission’s Proposal
