The Copyright Amendment Bill 2025 is the culmination of two years of consultation aimed at fixing a long-standing issue with orphan works, expanding online education exemptions, and making minor technical updates.
Division 2AAA in the Copyright Act 1968 limits remedies against those using copyrighted works when the owner can’t be found after a “reasonably diligent” search. While we still need clearer industry guidance on what qualifies as “reasonably diligent,” modern tools like TinEye, web searches, and AI make checking copyright straightforward.
This bill introduces requirements for copyright owners to issue a takedown notice before seeking a charge. Crucially, any claim must now be reasonable. I’ve seen copyright trolls demand over $1,000 for minor, low-traffic uses of dormant works. This new requirement gives small businesses leverage to negotiate and provides courts with clear guidance to lower litigation costs.
This is a major win for our galleries, libraries, museums, and educational institutions. With up to 70% of some collections categorised as orphaned, this bill grants them the freedom to finally share these works.
However, this bill completely misses the elephant in the room: AI training and data scraping.
Under current law, using copyrighted material to train AI generally requires a license, and reproducing substantial parts of works in AI outputs is an infringement. The real danger is how search engines and AI platforms (ChatGPT being a primary offender) scrape copyrighted websites to generate direct answers. By doing so, they eliminate the need for users to visit the original creator’s site, effectively appropriating their work. This critical issue must be addressed in the very near future.
Although One Nation supported this bill, we will be watching closely to see how it performs in practice. If small businesses acting in good faith aren’t adequately protected from predatory claims, we will push for further measures.
Transcript
The Copyright Amendment Bill 2025 is the result of consultation over two years to solve a problem that’s existed around orphan works. Schedule 1 of the bill creates a statutory framework that enables lawful use of orphan works when the copyright owner cannot be identified or located despite a reasonably diligent approach. Schedule 2 makes clear that copyright exemptions for learning institutions teaching in person extend to lessons taught online. Schedule 3 makes minor technical changes and updates.
The orphan works scheme creates division 2AAA in the Copyright Act 1968 to limit remedies for copyright infringement where the owner cannot be found despite a reasonably diligent approach. It would be good to have that further guidance on what ‘reasonably diligent’ means. I do expect there will be guidance from industry and media around that definition when the bill commences. Using internet search and websites like TinEye or asking AI if something is copyrighted is not hard to do; it’s quite easy. The act already limits backdated claims to six years; where the use predates that, only the last six years can be claimed as copyright infringement.
The bill requires a copyright owner who has detected use of their material to give the offending use a notice to take down the material. If they do, a charge does not result. It’s the copyright owner’s discretion. Any charge they do claim must be reasonable—there’s that word again. There’s a real issue with copyright trolls, who buy up dormant copyrights, look to see who has used that writing, photo or art then send a copyright claim. I’ve seen many examples of copyright trolls making claims above $1,000 for insignificant use such as a photo on a website that attracts little to no traffic. This legislation does introduce words like ‘reasonable’ to describe the charge. That is new, and it should help in these cases, because it gives a small business a chance to negotiate the charge. It may stop claims being made for minor use when no quantifiable benefit has resulted. The legislation provides the courts with clear rules and guidance that will reduce the cost of litigation for both parties.
One Nation will watch how this legislation works in practice. If small businesses are not being properly protected for in-good-faith breaches, further measures may be necessary. The good news is that this bill will allow galleries, libraries, museums and education institutions greater freedom to use works they have on hand. With an estimated 70 per cent of some collections being orphaned, this is a significant step forward.
The bill does not address the major concern about copyright law, which is using copyright material to train AI. Under the Copyright Act 1968, teaching and AI on copyrighted works generally requires permission or a licence from the copyright owner. Reproducing producing substantial parts of copyrighted works in an AI’s output, such as quoting long excerpts or reproducing poems or images is usually an infringement. The exception is a narrow fair-use exception for academic and news purposes. Where this becomes a problem is in areas of search where the old ten blue links in Google’s page of search results have been replaced with an AI answer, which uses information from a copyrighted website, generally removing the need to visit the site. Many artificial intelligence sites—ChatGPT being a major offender—will use data from a copyrighted site to answer a user question and even make recommendations for which website to use based on the data from a different site. The issue of AI appropriating copyrighted works or copyrighted webpages is an issue that will need to be addressed in the near future. One Nation will support this bill.
