The dialogue surrounding the rights of publishers in the age of artificial intelligence has taken an intriguing turn. Recently, Google has raised the idea that publishers should have the ability to opt out of having their works mined by generative AI systems. However, the specifics of how such a system would work remain undisclosed.
In a compelling twist, Google has suggested changes to Australian copyright law, aimed at allowing generative AI systems to scrape the Internet. The tech giant called on Australian policymakers to develop “copyright systems that enable appropriate and fair use of copyrighted content to enable the training of AI models in Australia on a broad and diverse range of data while supporting workable opt-outs for entities that prefer their data not to be trained in using AI systems.”
This call for a fair use exception is not unfamiliar ground for Google, but the concept of an opt-out option for publishers adds a fresh layer to their argument. When probed about how such a system might function, a spokesperson referred to a recent Google blog post. In it, the company suggested a dialogue around the creation of a community-developed web standard akin to the robots.txt system, allowing publishers to selectively opt out from parts of their sites being crawled.
The news has stirred the waters on a hot-button topic. Companies like News Corp have reportedly begun discussing payments with AI companies for scraping news articles. Meanwhile, OpenAI is currently embroiled in a lawsuit from authors Christopher Golden and Richard Kadrey, as well as comedian and writer Sarah Silverman, who are arguing that ChatGPT has been trained on their writing without their permission, or any financial compensation.
In a notable perspective, Toby Murray, associate professor at the University of Melbourne’s computing and information systems school, indicated that existing frameworks like Creative Commons might already provide the necessary groundwork. However, the underlying intentions of Google’s proposal might lean towards shaping early norms around non-payment for content.
The issue of copyright infringement, particularly affecting non-powerful entities, adds another complex layer to the discussion. The wider implications are yet to be fully explored, but some believe that smaller content creators could be adversely affected. With little bargaining power, these smaller content creators will be at the mercy of government regulators to protect them.
Google’s move towards an opt-out option makes sense, but it remains unclear what the implications of opting out would actually do to content creators. If Google is unable to leverage content to inform its AI models, then Google may look to penalise or remove sites from Google search results or AI-powered answers in its new search platform. This could be detrimental to content creators, no longer making this a viable financial option.
With the closure of submissions to the AI consultation last week, and the anticipation of hundreds of entries, the conversation around AI, copyright, and the balance of power between creators and technology firms is far from over. The early signs are that this will be a pivotal debate, positioning AI’s growth against the rights and interests of publishers, in a landscape that continues to evolve and challenge traditional boundaries.