What might generative AI mean for the IP rights system? How might prompt-based generation of novel content, including images and text, affect the IP rights administered by IP Australia?
Based on interviews, research and exploration of the capabilities of this new technology, the ventures team within IP Australia has made some preliminary observations. Several implications have been identified to help inform any future conversations about the topic:
- Generative AI is a step-change for the IP system – generative AI will introduce a host of new and profound questions, or amplify existing ones to a whole new level
- Generative AI will mean more actors in the IP system, making it a different system – by reducing the barrier to creating novelty, generative AI lowers the barrier to entry to the IP system. This is likely to have a material effect on how the IP system operates and how it is administered
- Generative AI will have cross-cutting effects across all parts of business – generative AI will intersect to varying degrees with the purpose, process and function of each of the four IP rights. Likewise, generative AI will be relevant across the functions of IP Australia and that of its customers and stakeholders
- Generative AI is not yet a mature technology – generative AI holds the potential for significant advancements, and relying on assumptions about what generative AI can't do based solely on its current performance is risky
- Norms around generative AI will need to evolve quickly – how, when and to what extent generative AI can and should be used within the context of the IP system will require establishing or adapting norms, expectations and responsibilities and gaining collective clarity around what's appropriate or not
- The impacts of generative AI will continue and evolve over time, rather than being a one-time transformation – the current direction of generative AI technology means that it will raise fundamental questions about the purpose, functions and processes of the IP system. This may escalate over time.
These observations underpin three core implications for the IP system regarding generative AI:
- Paradigm shift #1 – From inventing with tools, to tools inventing - generative AI means that we can no longer rely on the default assumption that new content (including ideas and inventions) are solely the result of human exertion
- Paradigm shift #2 – From scarcity to abundance - generative AI potentially changes the limiting factor for innovation, and moves it away from novelty. It will lead to more content, more novelty (or more content that is seemingly novel enough that it cannot be assumed to not be novel), easier access to the creation of novelty, more actors, and greater speed and capability for those actors
- Paradigm shift #3 – A fit-for-purpose system - together these raise significant questions and suggest that a fit-for-purpose IP rights system may need to look differently than it to what it currently does. At the very least they suggest it's risky to assume that the current approach is fit-for-purpose.
These observations are accompanied by a series of ‘provocations’ or scenario pieces, that illustrate different possibilities of how generative AI might intersect with each of the IP rights. These aren't forecasts or predictions; rather they are an attempt to test assumptions and reduce uncertainty about what could happen with generative AI.