A reference group focusing on copyright and artificial intelligence has been created by the federal government to tackle one of the largest issues emerging with the rapid growth of generative AI technology.
Attorney-General Mark Dreyfus on Tuesday announced the creation of the reference group, which will assist the federal government in better preparing for future copyright challenges emerging from the increasing usage of AI.
Dreyfus has hosted four roundtables on copyright this year, with the fourth taking place on Monday.
Through this process, the government has heard from 50 peak bodies and other organisations on copyright issues, with AI emerging as a key problem to address in the future.
“AI gives rise to a number of important copyright issues, including the material used to train AI models, transparency of inputs and outputs, the use of AI to create imitative works, and whether and when AI-generated works should receive copyright protection,” Dreyfus said in a statement.
The reference group will provide an ongoing mechanism for engagement with key stakeholders regarding copyright, including in the creative, media and tech sectors.
“Engagement with a broad range of stakeholders and sectors will help Australia harness AI opportunities, while continuing to support the vitality of our creative sector,” Dreyfus said.
The launch of OpenAI’s ChatGPT tool a year ago kickstarted a huge growth in the use and popularity of generative AI, with the app boasting 100 million users after just one month.
The skyrocketing popularity of generative AI has created copyright chaos, with several lawsuits already in progress.
A key concern centres on the usage of copyright material to train large language models used in generative AI tools, and the fact this can assist these tools to become direct competitors of the very copyright holders it has been trained on.
In September nearly 185,000 pirated books were discovered in a widely used dataset that was reportedly used to train Meta’s open source large language model, LLaMA.
A number of Australian authors, including Booker Prize winner Richard Flanagan and John Marsden, discovered that their books had been pirated in this dataset.
Flanagan labelled this the “biggest act of copyright theft in history”, while Marsden said it could lead to a “frightening and horrifying kind of tsunami of imitations which would do incredible, incalculable damage to the creative powers and efforts of human beings”.
In the US, the Authors Guild along with 17 well-known authors, including Jodi Picoult and George RR Martin have filed a lawsuit against ChatGPT for what they allege are “flagrant and harmful infringements” of copyright.
Getty Images launched a lawsuit against Stability AI earlier this year alleging that the generative AI tool copied and processed “millions of images protected by copyright and the associated metadata owned or represented by Getty Images absent a license”.
The New York Times has been mulling a potentially significant copyright challenge against ChatGPT, too.
The news giant this year updated its terms of service to prevent AI companies from using its content to train their models.
The company is also reportedly considering a lawsuit against ChatGPT for the use of its articles in its large language model in order to “protect the intellectual property rights” of the media firm.
If it goes ahead, this case could be the “most high profile legal tussle yet over copyright protection in the age of generative AI”, and could result in OpenAI being forced to wipe ChatGPT’s dataset and start again using non-copyright content.
The main concern of the New York Times is the use of its content to train ChatGPT and the potential then for the AI tool to become a direct competitor to it.
There have also been ongoing concerns over whether AI-generated content can enjoy the same copyright protections as content created by humans.
A US court earlier this year ruled that AI-generated images or creations cannot be copyrighted as “human authorship” is required for this legal protection.
This ruling potentially has a significant implication for the wealth of AI-generated content now being produced every day.
In contrast, a China court last week ruled that an AI-generated image can be covered by copyright law, as Semafor reported.
The Beijing Internet Court ruled that an image created with US-based text-to-image generator Stable Diffusion is protected by copyright.
This came after the creator sued a blogger for using the image without permission.