Australia may let AI companies scrape everything you post online, by default
The Albanese government is weighing changes that would give US tech giants open access to Australian content. Here is what that means for anyone who has ever posted, written or created anything on the internet.
Key points
- The Albanese government is actively considering reforms that would let AI companies scrape Australian online content without asking permission first.
- Everything posted publicly online, including social media, creative work and personal writing, could be used to train AI systems by default.
- Big tech companies have been lobbying the Australian government directly to make this kind of data access legal.
- No final decision has been made, and the reforms remain under review.
Right now, Australian copyright law is murky enough that an AI company training its models on something you wrote isn't clearly legal. These reforms would change that, tipping the scales firmly toward the tech companies.
Think of it this way. Every recipe you've shared in a Facebook group, every short story posted to a forum, every product review, every photo caption: all of it could become training data for a large language model (the technology behind chatbots like ChatGPT and Claude) without you opting in or being asked. We first reported on this policy direction on 15 September, when it emerged the government was pursuing a compromise with US tech giants that local artists weren't happy about.
What is actually being proposed?
The default position would flip. Instead of companies needing permission to use your content, you'd need to actively opt out if you didn't want it used.
As reported by The Guardian, Labor is facing pressure from US big tech to formalise this kind of access, and under the current thinking that pressure may well succeed.
For a nurse writing a health blog, a teacher sharing classroom resources, or a small business owner posting product descriptions, the upshot is the same. Your work feeds someone else's AI product, for free, unless you know to say no and know how to say it.
Should you be worried?
That depends on what you create. If you produce original writing, journalism or code for a living, this matters a great deal. Your output has commercial value, and these reforms could allow that value to be captured without compensation. The legal stakes aren't hypothetical: our 6 September story covered two US news outlets suing OpenAI and Microsoft and demanding AI models trained on their journalism be destroyed.
If you're a casual social media user, the personal impact is lower, but it isn't zero. Your posts still shape how AI models learn to write and sound like real people.
There's also a privacy angle. Even publicly posted content can contain personal details: your suburb, your health worries, your family situation. Once that data is absorbed into a training set, there's no pulling it back out.
No decision is final yet, which means there's still time for public submissions to shape the outcome.
Common questions
Does this affect content behind a login or paywall?
The reforms, as described, appear aimed at publicly accessible content. Material behind a paywall or a private account setting wouldn't automatically be in scope, but the detail of any final legislation will matter enormously.
Can you opt out?
A formal opt-out mechanism is part of what's being discussed. The critical problem is that opting out would be the user's responsibility, not the company's. Most people will never know they need to do it.
