Innovation and Commercialization Working Group

Partnership on AI (GPAI) for 2024, is committed to promote collaborative AI among GPAI Members by supporting projects towards promoting equitable access to critical resources for AI research and innovation, in compliance with applicable intellectual property protections and data protection legislation. As part of this commitment, India had proposed during the GPAI Summit 2023, two new thematic areas, which were included in the GPAI agenda: Sustainable Agriculture Collaborative AI for Global Partnership — May 12, 2025

The Innovation & Commercialization Working Group (I&C WG)’s mandate is to study and recommend tools and methods for driving international collaboration on AI R&D and innovation; advance research results into products and processes, and transfer these results to industry. Additionally, the I&C WG’s projects all have a special focus on SMEs due to their importance in the workforce. The I&C WG is comprised of 33 GPAI Experts, 13 External Experts, 2 Observers, and 2 GPAI Junior Investigators with varied backgrounds and sets of expertise that actively contribute to the I&C WG projects. Of these 33 Experts, 18 Experts come from an “Industry” background, 10 from the “Science” field, 3 with roles in the “Government”, 1 from “Civil Society”, and 1 from an “International Organization”. Among the 14 External Experts, 7 come from Science backgrounds, 6 have Industry backgrounds, and 1 is coming from an International Organization. The 2 Observers are from International Organizations, and both of our GPAI Junior Investigators come from an Industry background. Furthermore, the External Experts were nominated to the I&C WG by the Co-Chairs with the purpose of actively contributing their unique sets of expertise to the I&C WG’s projects. Their valuable experiences not only enrich the projects with their knowledge and network, but also reinforcing geographic diversity. The Observers also play a unique role in the I&C Working Group, notably contributing insights from their day-to-day role in their respective organizations. The 2 GPAI Junior Investigators in the I&C WG are individuals who are pursuing a PhD alongside of working in the respective organizations of GPAI Experts. Of the 29 GPAI Members, 22 of them are represented among the GPAI Experts in the I&C WG. In addition, among the GPAI Experts and the External Experts, a total of 25 countries are represented, spanning across the Americas, Africa, Asia, and Europe. Finally, GPAI considers gender diversity an important factor and actively strives to reach equal representation among the GPAI Experts. In 2023, the I&C WG was comprised of approximately 31% female GPAI Experts and 69% male GPAI Experts.— May 12, 2025

Experts in the agriculture and farming (A&F) sector hold considerable relevance for the GPAI Members in economic terms. Although A&F represents a low GDP percentage for some of the Members, the majority of the GPAI Members rank within the top 20 global list of exporters of food and agricultural products. What this means is that it is of utmost importance that we optimize the yields of this sector as demand for A&F products is likely to increase because of population growth and the reduction of arable land in the next 20 years. Furthermore, the A&F sector deals with issues that directly affect humankind’s survival. A top priority is the optimization of natural resources such as water, which is an area that needs to be addressed from an AI perspective. There is growing consensus within the scientific community regarding the growing need to build and maintain a drinkable water grid worldwide and the optimization of water resources deployed in A&F, especially in countries suffering from climate over-heating.— May 12, 2025

Regulating artificial intelligence (AI) presents one of the major challenges that governments, industries, and societies-at-large will have to tackle in both the present and the coming years. On one hand, AI will continue to affect every aspect of our lives in the future, which makes the regulation of AI most impactful on how we will interact with AI-powered applications. On the other hand, AI is advancing at unprecedented speed, which makes it challenging to foresee future directions and to align rather stable regulation with dynamic innovation in a reasonable way. As AI improves and diffuses, GPAI Members’ governments are considering the role of regulation to limit potential harms but not to hinder innovation, whether the regulation is legally binding (e.g., EU’s proposed AI Act) or non-binding (e.g., Singapore’s Model Governance Framework, Japanese METI’s AI Governance Framework). To help guide governments in this task, the I&C WG laid out a number of principles and best practices for a private session on AI regulation for the GPAI Summit in November 2021. In 2022, the I&C WG continued to finalize the principles for AI regulation and prepared the categorization of existing regulatory approaches. In 2023, the Regulation Project Advisory Group (Regulation Project) aimed to continue their effort to collect practical examples of how countries, institutions or other organizations support the industry to innovate while complying with AI regulation. Moreover, the Regulation Project aimed to outline concrete regulatory approaches and impact measures. The initial plan for 2023 focused on creating metrics to evaluate the impact of regulation on innovation, mapping regulatory approaches, and developing a library of resources. However, the project faced unique challenges due to the specific global environments, the dynamic nature of the work with evolving regulations, and the need to engage a diverse range of participants. This led the Regulation Project Advisory Group to change gears midway in the year and redirect our path for 2024. Ultimately, the I&C WG aims investigate the various AI regulations procedures deployed world-wide in the context of innovation. Its goal is to devise a collection of tools or practices to measure the impact of such procedures on innovation and commercialization. It will then gather, as best as possible, the value of the proposed frameworks, resources or indicators for the listed regulation procedures. The I&C WG does not work on standards or norms. Moreover, the I&C WG does not intend to propose any recommendation on what the “best” regulation policy is. But we do intend to propose ways to measure the impact of regulation on innovation and commercialization. This does include examples from low- and middle-income countries within the reach of the WGs experts to create a diverse set of practices. The Regulation Project, as well as the I&C WG-at-large, build on work from the OECD and actively looks for integration of OECD results into the GPAI work to ensure consistency and strengthened collaboration. — May 12, 2025

Small and Medium Enterprises (SMEs), Non-Profit Organizations (NGOs), and government-funded institutions typically employ 80% of a country's workforce. However, these organizations usually have low AI maturity as they often lack the resources, skills, data, or IT infrastructure to develop and adopt AI solutions. Understanding AI's capabilities, identifying AI use cases, and applying AI solutions are critical success factors for these organizations and the economies that host them. Early adopters of AI will have a critical advantage from experience with AI applications. The GPAI “Broad Adoption of AI by SMEs” working group has been working hard to address this issue and promote the widespread adoption of AI. This report highlights the significant achievements made in 2023 towards realizing this vision. The SMEs Committee recommendations were delivered as both a report, and also actualized as a web portal (“AI4SME Portal”) since 2021 which implements the report’s recommendations. The portal is meant for matching SMEs with low AI maturity and solution providers that have the relevant experience in delivering the AI solutions that meet the needs of the SMEs.— May 12, 2025

The Intellectual Property (“IP”) Advisory Committee of the Innovation and Commercialization Working Group (the “Committee”) continued its work to provide a platform to help stakeholders develop standardized contractual terms to facilitate artificial intelligence (“AI”) data and model sharing. The Committee hosted two hybrid multi-stakeholder workshops, one at the Max Planck Institute for Innovation and Competition (the “MPI”) in Munich, Germany, and another at Duke University (“Duke”) in Washington, D.C, United States. The workshop concept notes and agendas are included in the Appendices. The workshops, which were undertaken in collaboration with MPI and Duke, built upon the Committee’s 2022 report, entitled, Preliminary Report on Data and AI Model Licensing (2022) (the “2022 Report”).— May 12, 2025

The Innovation and Commercialization (I&C) Working Group’s mandate is to: • Study and recommend tools and methods for driving international collaboration on AI R&D and innovation, • Advance research results into products and processes, • Transfer these results to industry, with a special focus on SMEs. It will examine support measures to facilitate these advances, including standards and norms, selfcertification etc. The I&C Working Group is comprised of 42 Experts, 11 Specialists, and 6 Observers with varied backgrounds and expertise that contribute to the WG projects. Of these 42 Experts, 19 Experts come from an Industry background, 16 from the Science field, 4 with roles in the Government, 2 from Civil Society, and 1 from an International Organization. Among the 11 Specialists, 6 are of Science backgrounds and 5 are of Industry backgrounds. Of the 6 Observers, 4 have Industry backgrounds, 1 comes from a Science background, and 1 comes from an International Organization. Both the Specialists and Observers were nominated to the I&C WG Co-Chairs to actively contribute to the projects. More specifically, the Specialists were nominated to contribute their expertise to the IP Committee and Agro Committee. Their skillsets are of great value to the I&C Working Group, as they enrich the projects with their knowledge and network, but also contribute to geographic diversity. The Observers also play a unique role in the I&C Working Group, notably contributing to the IP Committee. There are three Observers from France, two from Canada and one from the OECD. The IP Committee also calls for an array of IP experts with regional diversity to contribute to the project, including Specialists from France, the United States, and the World Intellectual Property Organization (WIPO). All GPAI Members nominated 1 or 2 Experts to the Working Group, thus ensuring geographical diversity. Finally, GPAI considers gender diversity, and has achieved it with approximately 36% female Experts and 64% male Experts making up the I&C Working Group.— May 12, 2025

This new project (project 3) of the I&C WG commenced in February 2022 with a co-chaired leadership team composed by Inma Martinez (Spain), Digital Pioneer and AI Scientist, and Daniela Rus (USA), Deputy Dean of Research, Schwarzman College of Computing at Massachusetts Institute of Technology. It derived from Project 1, Broad Adoption of AI by SMEs, to address the specific needs and sector dynamics of agriculture and animal farming, which in many aspects, continue to deploy outdated methods and labor intensive practices that make agriculture ineffective, unsustainable, pollutant, and of low-paid (economically unattractive), and unnecessary duress for its workers. The Agriculture industry is currently undergoing an incredibly transformational and innovative moment in history thanks to digitalization and AI-driven technologies. If the 20th century agricultural industrialization created the mass-production of crops and animal farming in order to feed a growing population, the 21st century is forcing the Agriculture sectors to address radical challenges such as the demand for specific products on a global scale and out of season, climate change, the erosion of arable land due to antiquated and ineffective practices, and the preservation of potable water. In addition, a growing reversal of human eating habits towards more plant-based produce is contributing to the need to reconsider animal farming under its existing practices, not just on the moral grounds of their welfare, but because of the exorbitant use of arable land needed to feed them and the greenhouse effects that eating grain rather than grass causes in many of them. 20th century models that sought cheap routes to tackle agriculture are today lambasting the food value chains and our need to create a sustainable, climate aware, green economy. Project 3 is documenting AI deployments in agriculture and animal farming in the last 5 years to present evidence in an easy-to-use web portal that will showcase the potentiality and benefits of AI across a wide variety of agricultural sectors, geographies and the industrial development stages of the various GPAI Members. The purpose of Project 3 is to incentivize independent farmers and agricultural co-operatives to deploy AI and complementary digital technologies with clear return on investment, and sustainability practices.— May 12, 2025

In a data-driven world, data access is key to develop digital products. The IP Committee of the GPAI Innovation & Commercialization Working Group has initiated this work to identify global recommendations applicable regardless of the country. Data scraping is the process of extracting content from a website and importing it on a computer. The content can be used to then be analyzed or fed into an artificial intelligence algorithm. In certain instances, content scraped from public-facing websites may be protected by copyright and will require a license or an exception. Given that there is no international exception to copyright for data scraping, jurisdictions have very different approaches to the matter. As examples, data scraping in the United States can be allowed under fair use, provided it meets the criteria. Additionally, the European Union has introduced new text and data mining exceptions to copyright1 . Japan allows data scraping for computerized technical analysis only. There is quite the uncertainty on whether data scraping is subjected to the authorization of the right holder in other jurisdictions. Taking into account the diversity of applicable laws, these guidelines were designed to provide general recommendations for data scraping. So, the first graphic intends to be thought of as global guidelines to explain what one should and should not do, in order to avoid any intellectual property (IP) issues while web scraping and training AI. The second graphics are thought to present different exceptions that may be applicable depending on the jurisdiction. — May 12, 2025

The purpose of the field test was to validate the usability of the AI4SME Portal (“portal”) in matching SMEs and solution providers (“SPs”). Essentially, it was an opportunity to solicit relevant feedback on the portal features and the process of managing the portal by administering usability feedback surveys for SMEs and solution providers. The field test was participated by France, Germany, Poland, and Singapore (“field test participants”). — May 12, 2025

The premise that access to data is key for data-driven innovation—including for the development of artificial intelligence (AI) systems and applications—is broadly recognized. Yet multiple technical, economic and legal challenges to barrier-free and responsible data sharing persist. It is acknowledged that the standardization of data-sharing agreements may mitigate or help overcome some of such challenges and thereby aid and foster innovation across sectors and jurisdictions. In particular, standard terms and agreements can enhance legal certainty and reduce transaction costs related to contract negotiations and formation. — May 12, 2025

This is a simple and practical guide to intellectual property (IP) for AI practitioners. The targeted audience is small and medium-sized enterprises (SMEs) who plan to develop or employ AI technologies, but it is also useful for anyone who is interested in AI. In this document, “AI” is synonymous to an information system that uses machine learning (ML) technologies in some part of the system. An AI system is a machine-based system that is capable of influencing the environment by producing an output (detections, predictions, recommendations, or decisions) for a given set of objectives. It uses machine and/or human-based data and inputs to: (i) perceive real and/or virtual environments; (ii) abstract these perceptions into models through analysis in an automated manner (e.g., with machine learning), or manually, and (iii) use model inference to formulate options for outcomes. AI systems are designed to operate with varying levels of autonomy. This document consists of five sections. Chapter 1 (What is Intellectual Property (IP)?) provides an overview of the concepts of intellectual property. Chapter 2 (IP Issues in AI) describes IP issues specific to AI. Chapter 3 (Difference Between Jurisdictions) is on legal rules, with a special focus on differences between different jurisdictions. Chapter 4 (IP Management for SMEs) has tips for SMEs on how to manage IP. Each section is concluded with a list of useful resources. The readers are encouraged to follow these links to further understand the topics. Finally, the last chapter, Chapter 5, is dedicated to FAQs. — May 12, 2025

The Innovation and Commercialization (I&C) Working Group’s mandate is to: • study and recommend tools and methods for driving international collaboration on AI R&D and innovation, • advance research results into products and processes, • transfer these results to industry, with a special focus on SMEs. The I&C Working Group will examine support measures to facilitate these advances, including standards and norms, self-certification, etc. The I&C Working Group is comprised of 35 Experts and 12 Observers with varied backgrounds and expertise that contribute to the WG projects. Of these 35 Experts, 17 Experts come from an Industry background, 14 from the Science field, 2 from Civil Society, and 2 with roles in the Government. Of the 12 Observers, 4 have an Industry background, 3 come from international organizations, 3 from the Science field, 1 from Civil Society, and 1 from the Government. The Observers play a unique role in the I&C Working Group. One is nominated by the Egyptian Government, one comes from the OECD, two come from UNESCO, and eight were nominated by the I&C WG Co-Chairs to contribute to the projects, notably the IP Committee. The IP Committee calls for an array of IP experts with regional diversity to contribute to the project, including three French, two Canadians, two Japanese, and an American. All GPAI members nominated 1 or 2 Experts to the Working Group, thus ensuring geographical diversity. Finally, GPAI considers gender diversity, and has achieved it with approximately 39% female Experts and 61% male Experts making up the I&C Working Group.— May 12, 2025

The SME Committee of GPAI Experts recognizes that one of the biggest hurdles of AI adoption, especially for SMEs, is getting past the initial inertia. Through discussions and industry experience sharing, the Committee identified three main challenges contributing to the initial inertia: difficulty in finding trusted AI solution providers, lack of knowledge to understand AI and identify AI use cases, and lack of guidance for AI adoption strategy. As a result, the SME Committee specifically developed an AI Solution Portal to address these challenges. The Solution Portal is comprised of shared materials and programs from across GPAI member states, which reflect the common goal and spirit among the Experts in the Committee: the sense of urgency to help SMEs be prepared for the new economy. Each member state will be able to adopt the portal for local use by changing the logo, color scheme, images, information, and layout. The Solution Portal proposes various means in helping SMEs to get started in employing AI. While AI Solution Providers can list their solutions on the platform, the Solution Portal admin must approve those AI solutions before they are published. This approval process ensures trusted and quality AI solution listings on the platform. In addition, other than the technicalities of AI solutions, the platform also ensures that all listed AI solutions are compliant with established ethical principles on AI usage, such as GPAI's values and the OECD Principles on Artificial Intelligence. This way, SMEs can focus on applying AI to enhance their business operations rather than getting bogged down with extraneous details.— May 12, 2025

In the Working Group on Innovation & Commercialization (WG I&C) of the Global Partnership on AI (GPAI), 30 experts with varied backgrounds and expertise from 15 countries collaborate to study and recommend practical tools, methods and actions that enable private actors and research organizations to drive international collaboration on AI innovation, to develop R&D outputs into products and processes, and to exploit these results for commercialization and widespread adoption This group is chaired by Françoise Soulié-Fogelman, Scientific Advisor of Hub FranceIA; and JeanFrançois Gagné, founder and CEO of ElementAI. Both have been members of the European Commission’s High-Level Expert Group on AI. — May 12, 2025

This is a simple and practical guide to intellectual property (IP) for AI practitioners. The targeted audience is small and medium-sized enterprises (SMEs) who plan to develop or employ AI technologies, but it is also useful for anyone who is interested in AI. In this document, “AI” is synonymous to an information system that uses machine learning (ML) technologies in some part of the system. An AI system is a machine-based system that is capable of influencing the environment by producing an output (detections, predictions, recommendations, or decisions) for a given set of objectives. It uses machine and/or human-based data and inputs to: (i) perceive real and/or virtual environments; (ii) abstract these perceptions into models through analysis in an automated manner (e.g., with machine learning), or manually, and (iii) use model inference to formulate options for outcomes. AI systems are designed to operate with varying levels of autonomy. This document consists of five sections. Chapter 1 (What is Intellectual Property (IP)?) gives a high-level overview of the concepts of intellectual property. Chapter 2 (IP Issues in AI) describes IP issues specific to AI. Chapter 3 (Difference Between Jurisdictions) is on regulations, with a special focus on differences between different jurisdictions. Chapter 4 (IP Management for SMEs) has tips for SMEs on how to manage IP. Each section is concluded with a list of useful resources. The readers are encouraged to follow these links to further understand the topics. Finally, the last chapter, Chapter 5, is dedicated to FAQs— May 12, 2025

January 28, 2021