According to experts, the recent proposal by the tech company regarding generative AI has the potential to disrupt Copyright law and disproportionately impact smaller content creators. Google has proposed that publishers should have the ability to choose not to allow generative artificial intelligence systems to mine their works. However, the company has not provided any specific information about how this system works.

Google, in its submission to the Australian government’s review of the regulatory framework concerning artificial intelligence (AI), has proposed modifications to copyright law. The suggested changes aim to enable generative AI systems to access and extract data from the internet.

The company has issued a request to Australian policymakers to advocate for copyright systems that facilitate the suitable and equitable utilisation of copyrighted content. This would enable the training of artificial intelligence (AI) models in Australia using a wide and varied range of data. Additionally, the company suggests the implementation of feasible opt-out provisions for entities that do not wish to have their data utilised for AI system training purposes.

The company has previously conveyed its support for a fair use exception for AI systems to the Australian government. However, Google has recently introduced a novel argument advocating for an opt-out option for publishers.

In response to inquiries regarding the functionality of this system, a representative referred to a recent blog post published by Google. In the post, Google expressed its intention to foster a dialogue concerning the establishment of a web standard developed by the community. This proposed standard would resemble the existing robots.txt system, which grants publishers the ability to exclude specific sections of their websites from being indexed by search engines.

Google’s remarks are in response to reports that news organizations, including News Corp., have begun discussions with AI companies regarding compensation for the use of news article scraping.

According to Dr. Kayleen Manwaring, a senior lecturer at UNSW Law and Justice, generative AI systems will face significant challenges in the near future due to the issue of copyright.

In order to generate meaningful results, it is generally accepted that a substantial amount of data points, typically in the millions, is required. However, this necessity for a large dataset inevitably leads to the replication of copyrighted material, which can be considered a violation of numerous individuals’ copyrights.

According to Manwaring, there are variations in laws across different countries concerning the permissible content for AI systems. However, he argued that implementing an opt-out system would fundamentally challenge the principles of copyright.

In order to replicate copyrighted material, it is necessary to obtain the explicit permission of the copyright holder, rather than relying on an opt-out approach. The proposed suggestion entails a comprehensive overhaul of the existing exception-handling mechanism.

According to Toby Murray, an associate professor at the School of Computing and Information Systems at the University of Melbourne, Google’s proposal would require content creators to explicitly state whether their content could be absorbed by AI systems. However, he pointed out that current licencing schemes like Creative Commons already provide a means for creators to indicate the permissible usage of their works.

According to the speaker, there is a possibility that they are aiming to establish early norms that exempt other companies from paying for this content.

According to Manwaring, the breakdown of copyright could potentially occur if the issue remains unresolved, thereby potentially disadvantaging smaller content creators.

The ongoing concern of copyright infringement is expected to persist, especially as prominent entities continue to experience unauthorised use of their copyrighted material. Currently, it is highly probable that less influential entities are experiencing frequent instances of copyright infringement. This assumption is based on the widespread belief that AI training sets extensively incorporate content sourced from the internet.

During the Senate estimates session held in May, Senator Sarah Henderson, representing the Liberal Party, inquired with the communications department about the potential implementation of a scheme akin to the news Media Bargaining Code. The purpose of such a scheme would be to compel AI companies to pay for the act of scraping websites.

The department provided a response in July, in which it referred to the government’s consultation on AI regulation. It stated that the government is currently evaluating future policy settings for the news media as part of its news media assistance program. Additionally, the department mentioned that the government is contemplating a Treasury review of the News Media Bargaining code.

The deadline for submissions to the AI consultation was last week. It has been acknowledged that a considerable number of submissions have been received, however, as of now, none of them have been published on the internet.