Big Tech companies are engaging in intellectual property theft on a grand scale, spending squillions on AI. How can artists resist them? Josh Barnett reports.
When Atlassian co-founder Scott Farquhar appeared on ABC’s 7.30, he urged Australia to embrace the “opportunities” of Artificial Intelligence, including reforms to copyright law that would allow tech companies to mine creative works without permission.
For musicians, writers, and artists, the pitch is as familiar as it is ominous. Big tech companies promise innovation, but the fine print often also means extracting value from others’ labour without proper compensation.
Big Tech and your data
Silicon Valley treats our personal data: collects it en masse, monetises it, and leaves users with little control.
A Productivity Commission report warns that AI companies are attempting the same manoeuvre with creative content, sucking in vast datasets of music, images, and text, usually without consent or payment.
Creative Australia, the government’s peak arts body, has released a set of principles to guide how AI is used in the creative industries,
“For generative AI applications designed to produce creative outputs, these datasets are drawn from existing creative work which have been used predominantly without permission and with no recognition or remuneration for the original creator.”
If history is anything to go by, artists have every reason to be sceptical. Recently, Australia’s privacy regulator dropped a case against Clearview AI, despite the company scraping billions of personal images without consent.
Australia’s privacy regulator drops face recognition case against predatory Clearview AI
Music streamer Spotify has been repeatedly caught hosting AI rip-offs that impersonate real artists and siphon their royalties.
Spotify’s system allows scammers to pump AI-generated tracks directly into real artist profiles, diverting income away from musicians while the platform shrugs it off.
Ghost artists, fake playlists, and algorithm-friendly filler have become business as usual, a decade-long playbook of replacing real musicians with cheap, royalty-free slop, and now they don’t even have to pay the middleman.
This is the reality behind the rhetoric. Farquhar might talk about “opportunity,” but the track record of Big Tech shows a pattern: exploit first, deny responsibility later, and leave artists to pick up the bill.
Spotify’s Dark Secret: AI Music Is Stealing from Real Artists | The Soundcheck
Artists rights
Australia’s performing rights system is simple. If a song is played on the radio, streamed in a café, or blasted through a shopping centre, APRA AMCOS ensures royalties flow back to the songwriter. If you use creative work, you pay for it.
Why should AI be different? If a model is trained on Midnight Oil’s catalogue or Kylie Minogue’s hits, the artists deserve to be paid. Anything else is theft disguised as ‘fair use’.
For visual artists and authors, Copyright laws offer protection, but as Farquhar points out, these laws are no longer fully fit for purpose and need updating.
Big revenue, but even bigger costs
OpenAI recently announced it reached $10B in annualised revenue, doubling from about $5.5B the year before. Yet behind that headline growth lies an uncomfortable truth: running advanced AI like ChatGPT is extremely expensive. Unlike traditional software, where serving more users adds negligible cost, every AI query consumes costly compute power.
Training AI models also requires expensive, specialised hardware and consumes a significant amount of energy. Even after the $10B revenue mark, OpenAI still “burns billions more” a year on talent and infrastructure and hasn’t disclosed profits. Online newspaper TechCrunch reports OpenAI’s losses last year near $5B.
This breakneck spending isn’t unique to OpenAI: industry-wide, planned AI infrastructure investment is enormous. Microsoft alone may spend over $80–110 billion on AI data centres next year, even though total generative AI revenues are still under $10 billion.
AI infrastructure also has a short shelf life. High-performance server clusters run hot and hard. A Google AI architect noted that under heavy load, data-centre processors might survive only 1–3 years before needing replacement.
The rug pull
The economics of the AI boom point to an eventual rug pull. After luring hundreds of millions with powerful AI tools, companies will have to change the deal through price hikes, reduced free access, or service cutbacks. OpenAI is reportedly eyeing a jump in ChatGPT’s subscription fee to about $44 a month, while others roll out pricier “priority” tiers and strip back free usage once growth slows.
Some analysts warn the AI boom looks like a bubble. Without major efficiency gains, today’s gold rush built on costly, short-lived infrastructure could end with services becoming too expensive, with uptake slowing down, or collapsing altogether.
The threat to artists is still real, though, and government regulation is required.
We’ve heard it all before, trust us, innovate, move fast. And time after time, Big Tech have shown they can’t be trusted to have consumers’ interests at heart.
Look at the crypto boom, the hype around NFTs, the rise and fall of social media giants, and Google’s long track record of avoiding its tax obligations in Australia.
If Google won’t pay its fair share on billions in profits, why should we let it take our creative work for free, build a new product out of it, and then ship the profits offshore?
Creative Australia’s principles are clear enough: pay artists, demand transparency, and put creatives in the room when policy is being made.
Anything less, and we’re not building an AI revolution; we’re just handing over our cultural identity to a cartel of global tech monopolies.
Josh is a professional musician and cameraman who is now working with Michael West Media to develop The West Report and other visual content across major social media channels