According to Google, publishers should have the option to opt out of having their works mined by generative artificial intelligence systems. However, Google has not provided details on how such a system would function.
In its submission to the Australian government’s review of AI regulations, Google stated that copyright law should be modified to allow for generative AI systems to scrape the internet.
The company called for Australian policymakers to support “copyright systems that enable appropriate and fair use of copyrighted content to facilitate AI training on a diverse range of data, while also providing opt-out mechanisms for entities that do not wish to have their data trained using AI systems.”
While Google has previously expressed its support for a fair use exception for AI systems to the Australian government, the concept of an opt-out option for publishers is a new argument.
When asked about the workings of such a system, a spokesperson referred to a recent blog post by Google that discussed the possibility of creating a web standard, similar to the robots.txt system, which allows publishers to opt out of having certain parts of their websites crawled by search engines.
Google’s comments come as news companies like News Corp have reportedly begun conversations with AI companies regarding compensation for scraping news articles.
Dr. Kayleen Manwaring, a senior lecturer at UNSW Law and Justice, highlighted copyright as a major challenge for generative AI systems in the future. She explained that generating useful outcomes often requires millions of data points, which involves copying and potentially breaches copyright laws.
Toby Murray, associate professor at the University of Melbourne’s computing and information systems school, stated that Google’s proposal would place the responsibility on content creators to specify whether AI systems can use their content. However, he pointed out that existing licensing schemes like Creative Commons already enable creators to define the permissible use of their works.
Manwaring warned that if the copyright issue is not resolved, it could undermine the rights of smaller content creators. She expressed concerns that powerful entities may have their copyrights infringed, while smaller entities could face widespread infringement if AI training sets extensively use internet material.
During Senate estimates in May, Senator Sarah Henderson inquired whether the government was considering a scheme similar to the news media bargaining code to ensure that AI companies pay for scraping sites. In July, the communications department responded, stating that the government was examining future policy settings for news media and considering the Treasury’s review of the news media bargaining code as part of its news media assistance program.
The AI consultation period closed last week, and hundreds of submissions have been received. However, none have been published online thus far.