Sydney: Prime Minister Anthony Albanese outlined the government’s response to local copyright and data center issues in a speech at the University of Sydney yesterday. The Australian government is trying to draw a strong line of defense against AI infringements of physical and intellectual property.
This is new territory for Australian legislation, and is in the face of massive AI uptake and investment in Australia. This is also the first clear indication of the government’s attitude. It comes after strong lobbying for arts protection and the rise of many local issues raised by data center developments.
The next big, unavoidable issue is copyright. Australian copyright law is an average international standard and equally vulnerable to data theft and AI hacking. It was never designed to manage AI phenomena. There was harsh criticism for the lack of action to regulate AI.
It even allegedly lobbies to “legalize copyright infringement” and blur IP rights. The government will inevitably have to codify and decide on these issues. Without statutory law, courts can only do so much, and they can’t do much.
Australia is a test case
The rest of the world has the same problems, but Australia is unique as a test case for managing AI. There is a large market overlap with international IP. Australian-made products are marketed outside Australia under license and copyright material is published and distributed overseas.
Therefore, these intellectual properties of very high value must be protected. AI effectively reduces the creative process, devaluing IP. AI products in turn also compete in the same markets.
Australia’s big advantage is being able to start from scratch managing these issues. There is no specific existing legislation or regulation on AI. An entire regulatory system can be built and tested for effective application.
Regulation of AI against existing laws
Existing copyright laws help to a point, but not enough:
Copyright materials are property and their ownership is not negotiable. Ownership is attributed to a legal entity by default. Property rules do not need to be “fixed” but enforced.
UA is not a legal entity in the same sense as a natural person or corporation. Cannot own or hold property.
Any use of copyrighted materials is covered by existing laws. If you want to use it, buy the right to use it, whether it’s a textbook or a comic.
AI cannot claim to have created IP. By definition, it derives what it produces from copyrighted material. This is the exact opposite of ownership. It is direct proof of non-ownership.
The main problems come with “scraping” in the AI training course. This includes training LLMs on large amounts of data, much of it commercial copyright. This is usually done without the consent of the copyright owners or any compensation.
Content derived from scraping feeds back into the same markets that the training materials originate from. Artificial intelligence is “diluting” these markets directly. It’s generating a lot of largely useless AI bottlenecks at the expense of copyrighted material producers.
Markets against laws
The biggest problem with AI scraping is the lack of drive on the part of AI developers. They have resisted any recognition of basic copyright laws.
This whole situation could be completely avoided with definitive laws. In theory, all current copyright laws cover scraping simply because of the direct use of IP in scraping. In practice, none of these laws work at all and are not being enforced. AI anti-scratch test cases there are many, but so far there are no clear decisions. These endless disputes are just dragging the chain. It would be simpler and much cheaper to reach a workable agreement for copyright owners.
This is where Australia can get ahead of disasters and provide direct legal solutions. The alternative is a legal mess with incredibly slow turnaround times for decisions.
An important reality of AI that is being overlooked
AI LLM training is a big and expensive process, but it pays off in many ways. The values derived from this training extend beyond writing a book, doing one’s homework, and compromising the entire future of higher education.
The real value of AI training has little, if anything, to do with stealing creative content anyway. This training, especially at the language level, gives great efficiency. It provides linguistic fluency and dynamic ability defines the needs of AI. Training is critical to AI performance and future development.
By comparison, the value of creative content is simply training the AI to learn those skills. A price can easily be established and standardized for general use.
Should a textbook have a price? Yeah, so what’s the problem? Do you, or can you, master the content of the textbook? Only if you buy the rights. Should you own that content? No. You are simply paying for the use of the content.
The world must accept AI regulation
Australia is trying to step forward to clean up this legal mess and make it manageable. Copyrights are huge assets around the world. International laws must be on the same page to work.
AI disputes against copyright are getting nowhere and are very slow in this environment. Most copyright laws can simply be adjusted to define copyright ownership rights in relation to AI. It is time for AI to learn to respect property rights.
