Hello, World: Artificial intelligence and its use in the public sector by Jamie Berryhill, Kévin Kok Heang, Rob Clogher, Keegan McBride published by OECD/OPSI (11/2019).
“Artificial Intelligence (AI) holds great promise for the public sector, and governments are in a unique position in relation to AI. They are capable of setting national priorities, investments and regulations for AI, and can also use it to redefine the ways in which the public sector creates policies and services. Hype around emerging technologies often overstates or obscures practical applications. An understanding of AI is therefore critical for helping policy makers and civil servants determine whether it can help them achieve their missions.
Individuals and businesses are increasingly interacting with AI. Although the field has been researched and discussed for more than 70 years, there is still no uniformly accepted definition; AI means different things to different people. According to the 2019 OECD Recommendation on Artificial Intelligence, AI currently refers to machine-based systems that can, for a given set of human-defined objectives, make predictions, recommendations or decisions influencing real or virtual environments. The OECD Observatory of Public Sector Innovation (OPSI) (https://oecd-opsi.org) has developed this primer to help determine what AI means for public sector innovation, and to help public servants understand AI and navigate its implications for policies and services.
At a technical level, while AI takes a variety of forms, all AI today can be classified as “narrow AI”. In other words, it can be used for specific tasks, for example handling and interpreting text through natural language processing, detecting and classifying objects through computer vision, and recognising and interpreting spoken language and translating it into text through speech recognition. Approaches such as “unsupervised learning”, “supervised learning”, “reinforcement learning”, and “deep learning”, which sit under the umbrella of “machine learning”, hold significant potential for a variety of tasks, yet each has its own strengths and limitations. It is important to note, however, that every AI project starts from the same point: data. Governments must ensure they have access to sufficient, quality, unbiased data before they can fully and ethically take advantage of these techniques.
In adopting AI, the public sector has trailed behind the private sector. However, governments are seeking to rapidly catch up. To catalyse AI-driven innovation, an initial mapping conducted by the OECD identified 50 countries (including the European Union) that have launched, or have plans to launch, national AI strategies. While at different stages of development, these include some common themes: economic development, trust and ethics, security and enhancing the talent pipeline. Of these 50 countries, 36 have (or plan to have) either separate strategies in place for public sector AI, or a dedicated focus embedded within a broader strategy. This is critical, as it allows AI to be integrated into the entire policy-making and service design process. These public sector components often promote a number of common themes, such as:
· experimentation with, and sometimes funding for, government AI to automate processes, guide decision making and develop anticipatory services for citizens.
· cross-government, cross-sector and international collaboration through councils, networks, communities and partnerships.
· strategic management and use of government data, including open data, to fuel AI in all sectors…”
