Reports

AI language models are a key component of natural language processing (NLP), a field of artificial intelligence (AI) focused on enabling computers to understand and generate human language. Language models and other NLP approaches involve developing algorithms and models that can process, analyse and generate natural language text or speech trained on vast amounts of data using techniques ranging from rule-based approaches to statistical models and deep learning. The application of language models is diverse and includes text completion, language translation, chatbots, virtual assistants and speech recognition. This report offers an overview of the AI language model and NLP landscape with current and emerging policy responses from around the world. It explores the basic building blocks of language models from a technical perspective using the OECD Framework for the Classification of AI Systems. The report also presents policy considerations through the lens of the OECD AI Principles.

Digital markets have raised a number of new competition challenges. Ex-post competition policy appears not to be able to address them in their entirety and with the necessary speed. There is considerable consensus, among academics and policy-makers, that ex-ante regulatory policies are needed to avoid competition being stifled in these markets, with a negative impact on productivity and innovation. As a result, major OECD economies are discussing or have approved regulatory proposals with the aim to foster contestability and fair trade in digital markets.

The provision of key health technologies and products such as vaccines and antibiotics is insufficient in purely competitive and volume-based markets, requiring new revenue streams for sustainability. Recent developments in health innovation suggest that innovative collaborative mechanisms can be effective in addressing this issue. In the domains of vaccines and antibiotics, these approaches should incorporate shared research investment, long-term access planning, the provision of manufacturing infrastructure, supply chains, and financial returns. Collaborative approaches such as subscription models could be piloted at the regional level, while other models could be developed to delink innovation, manufacturing, and access from sales volume and revenue. Finally, blended finance instruments from the development field could encourage greater collaboration among established and emerging stakeholders in health innovation. These stakeholders should work together to create, test, access, and implement more collaborative approaches to health innovation to share upfront investments, mitigate risks of failure, and accelerate market access.

This report examines privacy-enhancing technologies (PETs), which are digital solutions that allow information to be collected, processed, analysed, and shared while protecting data confidentiality and privacy. The report reviews recent technological advancements and evaluates the effectiveness of different types of PETs, as well as the challenges and opportunities they present. It also outlines current regulatory and policy approaches to PETs to help privacy enforcement authorities and policy makers better understand how they can be used to enhance privacy and data protection, and to improve overall data governance.

Policy makers are increasingly grappling with the stability implications of global value chains (GVCs), as widespread supply shortages following the COVID-19 pandemic and the Russian Federation’s large-scale aggression against Ukraine have disrupted the economic recovery and contributed to high inflation. This paper provides a tool to assess vulnerabilities in GVCs by drawing a detailed map of dependencies based on new indicators constructed from the OECD Inter-Country Input-Output tables. The key findings are as follows. First, GVC dependencies increase with both the size of foreign exposures and the length of foreign value chains. Second, in some industries, such as the automotive and ICT industries, vulnerabilities from high GVC dependence are amplified by high geographic concentration of suppliers or buyers. Third, the People’s Republic of China is the most critical choke point in GVCs across a broad range of industries, both as a dominant supplier and as a dominant buyer.

The transition to climate neutrality requires cost reductions in existing clean technologies to enable rapid deployment on a large scale, as well as the development of emerging technologies such as green hydrogen. This policy paper argues that science, technology, innovation, and industrial (STI&I) policies focusing on developing and deploying low-carbon technologies are crucial to achieving carbon neutrality. It notes however that the current level of innovation is insufficient to meet the net-zero challenge due to a policy emphasis on deployment rather than research and development (R&D) support. The paper explores the rationale for more ambitious STI&I policies targeted at R&D for climate neutrality and provides policy recommendations for an effective innovation policy for net-zero, including its interaction with the broader climate policy package.

Artificial intelligence (AI) is transforming economies and promising new opportunities for productivity, growth, and resilience. Countries are responding with national AI strategies to capitalise on these transformations. However, no country today has data on, or a targeted plan for, national AI compute capacity. This policy blind-spot may jeopardise domestic economic goals. This report provides the first blueprint for policy makers to help assess and plan for the national AI compute capacity needed to enable productivity gains and capture AI’s full economic potential. It provides guidance for policy makers on how to develop a national AI compute plan along three dimensions: capacity (availability and use), effectiveness (people, policy, innovation, access), and resilience (security, sovereignty, sustainability). The report also defines AI compute, takes stock of indicators, datasets, and proxies for measuring national AI compute capacity, and identifies obstacles to measuring and benchmarking national AI compute capacity across countries.

This report provides representative, cross-country estimates of the artificial intelligence (AI) workforce across OECD countries. The AI workforce is defined as the subset of workers with skills in statistics, computer science and machine learning who could actively develop and maintain AI systems. For countries that wish to be at the forefront of AI development, understanding the AI workforce is crucial to building and nurturing a talent pipeline, and ensuring that those who create AI reflect the diversity of society. This report uses data from online job vacancies to measure the within-occupation intensity of AI skill demand. The within-occupation AI intensity is then weighted to employment by occupation in labour force surveys to provide estimates of the size and growth of the AI workforce over time.