Share

HomeOur work > Innovation and commercialization > Protecting AI innovation, intellectual property

Protecting AI innovation, intellectual property

Both the development and use of AI technologies have the potential to be hindered by several identified challenges, including the questions such as: how to efficiently protect investment through intellectual property (IP) protection within a company developing new AI technologies; how can a company’s data be protected; what kind of intellectual property rights (IPRs) will be created, and how will ownership of such IP be organized; and how can a company address the different jurisdictions of IPRs, and how could this deter innovation?


In meeting such challenges, a business must understand how IPRs are handled in different situations, and especially across different jurisdictions, and set an IP management strategy so that the company can optimize their IPs while keeping scalable benefits and IP-related risks minimum. A few factors make this task particularly difficult:

  • Current IPR protection mechanisms are very complex.
  • Different jurisdictions have slightly different rules and practices.
  • New concepts such as data, raw data, labelled data, structured data, datasets, natural language processing (NLP), and image processing, raise questions on how to interact with existing IP mechanisms and how to shape future ones to cover the gaps.
  • There are many options to choose an IP management strategy: e.g., for a new idea, file a patent, saving know-how, or publish.


This knowledge is particularly important for organizations that have limited resources in their legal and/or IP functions when negotiating with a more resourceful company.


The main objective of the project is to achieve a deeper understanding of the challenges that organizations face in the area of IP as it relates to AI. This project will be conducted from the knowledge of current effective IP laws, best practices, activities and mechanisms related to IP, the overview of how individual organisations are currently handling IPs and differences across geographies. This knowledge will be the basis for the creation of materials to help organizations manage IP more effectively and distribute them to a large audience — especially through a simple IP Primer, with a Q&A to be designed to be easily accessible for entrepreneurs. Furthermore, the following outcomes are expected:

  • Ensuring AI IPs are protected.
  • Helping businesses to be more confident in terms of IP management.
  • Reducing the negotiation-power gap between large companies and less-resourceful entities.

 GPAI intellectual property (IP) primer (November 2021)

Short-term objectives

Provide a guidebook for businesses that want to develop or use AI technologies and that innovate using AI technologies. This guidebook will help businesses understand the current IP landscape and optimize their business strategy, alleviating IP-related risks in AI innovations. Two main activities will be conducted in 2021 with results to be shown at the 2021 Summit:

  • Publication of the GPAI IP Primer v1 and distribution to GPAI members.
  • Conducting six case studies on how businesses are handling IP risks regarding AI-related innovations and commercialization.

Long-term deliverable

Activities to be considered in 2022:

  • Incorporate implications of EU Data Governance Act, Digital Marked Act, Data Base Directive, and Machine Directive.
  • Update the GPAI IP Primer based on the case studies and disseminate to wider audience as IP Primer v2.
  • Publish a second document, GPAI IP Expert, for enriching the debate at an expert level.
  • Explore a pattern of clauses of framework data access, sharing and flows for implementing it to Free Trade Agreements, electronic trade rules under WTO, and/or transnational regional agreements among like-minded partners.
  • Produce key recommendations for harmonization of rules and mutual recognition of patenting computer-implemented AI inventions.