1. A ‘Google Zero’ World?
“I got the new stuff!” are the background lyrics of the latest Google advertisement about the future of online search. “Overviews” is a new feature that is gradually being rolled out to Google’s users globally. This new specialised feature makes it possible to summarise the web and show users an overview of answers to search queries. According to Liz Reid, the newly installed Head of Search at Google, “It can take a bunch of the hard work out of searching, so you can focus on the parts you want to do to get things done, or on the parts of exploring that you find exciting.”
Not every search query is complex. But when your search query requires combing the Internet and providing a nuanced response, AI-powered overview functionality is likely to become increasingly prevalent. We are entering the “Google Zero” paradigm—a world where AI-powered search provides summaries of search requests that may not even reference the original sources of the underlying content.
The “overviews” functionality “sent shockwaves around the web.” What happens to the Internet when major tech companies no longer support references to content creators? What happens to information created by online newspapers, magazines, and bloggers if Google does not provide links to their content? What will “Google Zero” mean to online audience traffic flows (specifically, blogs and podcasts)? Would there be any incentives to create original content? And what is the fate of copyright law in this new world order?
2. Copyright & AI (Search) Agents
Copyright law has long been the cornerstone of protecting the intellectual property (IP) rights of content creators and incentivising creativity. It ensures that creators can benefit from their work, thus encouraging the production of new content. In the era of AI-generated summaries, the effectiveness of copyright in the digital space faces an existential challenge. If summaries are provided without proper attribution or compensation, the foundational principles of copyright could easily be undermined. As a society, we may be at the point of no return: without fact-finding and original (not AI-generated) content, democratic societies cannot function or exist.
We believe that AI is a net-benefit to humanity, but its implications have been largely misunderstood. The emergence of gen-AI requires that we revisit existing copyright norms and explore what new norms should be established. To better understand the copyright implications of these new AI-powered advancements, we need to address the concept of the “AI Assistant” that permeates this revolutionary transition towards the consumption of content on the “open internet.”
Many of these emerging AI solutions are presented as “AI assistants”, “AI copilots,” or “AI agents.” Such terminology raises numerous questions: Whether and how such AI-powered tools are different from other software-based solutions? Is such a distinction between “traditional” software applications and AI-powered tools justified?
From the perspective of AI companies, it makes a lot of sense to emphasize the emergence of a new category of social reality and new narratives of creativity. The introduction of the notion of “AI Assistants” in popular jargon could be seen as an attempt to suggest that the way people consume information online is evolving.
As dozens of copyright lawsuits filed in U.S. courts indicate, one of the crucial issues is whether online content and information, often behind paywalls or otherwise protected by IP, can be used to train large language models (LLMs) and whether AI companies can rely on the fair use doctrine. The argument proposed by the AI companies is that an generative AI models are not engaged in an IP infringement of any sort; rather, they use human knowledge in a transformative way and provide the users the opportunity to interact with the content in new ways.
If this trend continues, we will likely see a pushback from online content creators–individuals as well as companies. In our view, if original content is used to feed AI assistants, then content creators will most likely stop generating content. To avoid this kind of deadlock, a new technological infrastructure and a new social contract with technology become necessary. Finding an appropriate mechanism for remuneration must be front and centre of this project to ensure that the incentive to create remains and an equitable share of rewards flows to creators.
3. Creativity & Remuneration With, Between, & Via AI Agents
Here, we envision that a new framework for content creation and exchange is emerging. In this AI-powered framework, content creators and users will interact via AI agents. This rapidly evolving, AI-powered multi-agent universe allows all members to have representation through their own AI agents, which interact with one another. In this new architecture, each content creator can “reveal” their content through their digital agents while maintaining ownership and control over how this content is accessed and used. The exchange of information will happen between and via AI agents.
Today, content creators typically depend on information intermediaries to publish their work. In the future, individual creators will have the power to truly own their content by publishing it via AI-powered agents. This new paradigm involves two key components: (a) a technological layer consisting of personal data storage instances and a sophisticated, LLM-powered interface through which this information will be accessible, and (b) a new suite of rights management tools that will dictate the conditions under which the content can be accessed. Think of these rights management tools as super-charged, AI-powered IP rights management assistants, akin to Creative Commons licenses but tailored for an AI-driven interface. Representative AI agents will “negotiate” in the back end with regard to what information can be accessed and under what terms.
Content dissemination within the network of AI agents must be both information-rich and action-rich. In the near future, while meta tags and robots.txt files remain invisible to humans, they will empower AI search agents to comprehend the underlying content more profoundly. Natural language which we humans use in our communication is also becoming the default means for AI agents to communicate. LLMs will enable AI agents to prompt each other for deeper understanding, providing all relevant information based on the end user’s context. Just as today’s websites use meta tags, content searches with AI agents will involve LLM-powered meta conversations that will be action-rich and happen in microseconds. This transformation of search can be summarized in one phrase: “The search will be done for us, rather than by us.”
This future is not a distant dream but a rapidly emerging reality. Each of the authors of this paper already has their own “AI knowledge twin.” Our AI knowledge twins house our personal publications, which we make available to our social media followers and students through this new interface.
Figure 1: Content Creation and Dissemination With, Between, and Via AI Agents
From a technological perspective, “AI knowledge twins” operate on each owner’s data and are uniquely organised. These twins leverage machine learing (ML) algorithms, enabling them to learn from data, adapt to user preferences, and improve performance over time. These algorithms enhance the AI’s capacity to provide personalised and relevant insights. Crucially, natural language processing (NLP) is employed to comprehend user queries, process information, and generate context-aware responses.
AI knowledge twins are designed to understand the context of conversations and interactions. Rather than simply reading the context, users of AI knowledge twins—whether humans or other AI agents—can enjoy a personalised experience. These AI knowledge twins are fully controlled by their owners, who can adapt the tone, style, and complexity of responses. Additionally, they can interact with users in multiple languages, enhancing both accessibility and usability. Owners have complete control over their AI knowledge twins, including the ability to update or modify the underlying knowledge base by adding new information and adjusting its personality.
In the light of current technology trends, it is not difficult to imagine that in the near future, these AI twins will include an additional layer of access rights and controls. These controls will determine the specific conditions under which human users and AI-powered search agents can interact with the content provided by AI-powered knowledge agents.
If this is happens, perhaps we should no longer fear “Google Zero.” In the age of AI, information exchange will evolve into interactions with, between, and via AI agents, where creators will have more ownership and control over their content. Content owners will be able to specify their terms of use, requiring both human and machine users to agree to these conditions. In this way, AI disrupts and affects every stage of the creativity stack: the production, distribution, consumption, and rewarding of content.
Perhaps “Google Zero” marks a time when information is starting to be organised in new ways and when a new social contract with technology needs to be considered. The key point is not to limit or slow down development, innovation, or the utilisation of knowledge. Instead, it is about recognizing how creativity is constantly disrupted in real-time and ensuring that credit and value flow fairly to their rightful owners, and not to those who do not deserve it.
_____________________________
To make sure you do not miss out on regular updates from the Kluwer Copyright Blog, please subscribe here.
Kluwer IP Law
The 2022 Future Ready Lawyer survey showed that 79% of lawyers think that the importance of legal technology will increase for next year. With Kluwer IP Law you can navigate the increasingly global practice of IP law with specialized, local and cross-border information and tools from every preferred location. Are you, as an IP professional, ready for the future?
Learn how Kluwer IP Law can support you.
