Aussie music legends’ works ‘sucked into AI’ without consent, sparking urgent copyright reform calls

Australia’s most celebrated musical artists are having their life’s work, including iconic songs by AC/DC and Kylie Minogue, systematically ingested into massive artificial intelligence (AI) datasets without their permission or compensation. This concerning development, highlighted in a recent investigation by The Conversation AU, reveals a glaring deficiency in Australian copyright law, leaving creators vulnerable in the rapidly evolving digital landscape.

The findings underscore a growing global ethical and legal dilemma: how to regulate AI’s insatiable appetite for data, much of which is copyrighted material. For Australian musicians, the stakes are particularly high, with fears their creative output could be used to train AI models that generate new music, potentially devaluing human artistry and livelihoods.

Unseen Digital Harvesting of Aussie Hits

The Conversation AU’s report details how major AI developers have trawled the internet, indiscriminately scraping colossal volumes of data to train their algorithms. This digital dragnet has captured millions of copyrighted works, including the discographies of Australia’s most beloved performers. The sheer scale of this data collection means that virtually any prominent Australian artist – from Cold Chisel to The Wiggles – is likely to have their material embedded within these AI models. The problem extends beyond major acts, encompassing independent artists and emerging talents whose works, available online, are equally susceptible to this unconsented digital harvesting.

The core issue is that current Australian copyright legislation, designed for a pre-AI era, does not explicitly address the use of copyrighted material for AI training. While some exceptions for research or private study exist, the commercial scale and intent of AI training push these boundaries to breaking point, leaving artists with minimal legal recourse to protect their intellectual property or demand fair remuneration. The absence of clear guidelines creates a legal grey area that tech companies have been quick to exploit, prioritising innovation over creator rights.

Copyright Laws Lag Behind AI Advance

The current legal framework in Australia operates under an “opt-out” model, meaning creators generally have to actively discover their work has been used and then initiate complex, often costly, legal challenges to have it removed. This is a near-impossible task given the opaque nature of AI datasets and the vast quantities of data involved. Industry bodies are advocating for an “opt-in” system, where explicit consent is required before creative works can be used for AI training, ensuring artists retain control over their intellectual property and can negotiate fair licensing agreements.

Comparatively, other nations are beginning to grapple with this issue more proactively. The European Union, for example, has introduced provisions within its proposed AI Act that aim to mandate transparency around the data used for AI training and provide clearer avenues for rights holders to protect their works. Australia’s slower pace of reform risks leaving its creative industries exposed and disadvantaged on the global stage, potentially stifling local innovation in industries that rely on intellectual property protection.

Industry Calls for Urgent Legislative Reform

The revelation has ignited calls from across the Australian music industry for urgent legislative reform. Key organisations representing musicians, record labels, and publishers are now lobbying the federal government to update copyright laws to explicitly address AI training and data scraping. They argue that without strong legal protections, the future viability of the Australian music industry is at risk. Artists rely on their intellectual property to earn a living, and if AI can freely replicate or derive new works from their output without compensation, the incentive to create original music diminishes significantly.

Proposed solutions include implementing a mandatory licensing scheme for AI training data, similar to existing models for broadcast or public performance, which would ensure artists receive fair royalties. There are also calls for greater transparency from AI developers regarding the datasets they use, allowing rights holders to identify if their works have been included. The creative sector maintains that any framework must balance technological advancement with the fundamental rights of creators, ensuring that AI innovation does not come at the expense of human artistry and fair compensation.