AI models rely on data to function. The quality of data that gets fed into a model determines the quality of its outputs – a classic representation of the ‘garbage in, garbage out’ axiom. In fact, the lack of quality data may even exacerbate bias and discrimination, particularly against vulnerable groups – pushing them further behind. Therefore, the degree of accuracy, relevance, and representativeness of a data set will impact the reliability and trustworthiness of results and insights the data is informing.

Digital public infrastructure, as an interoperable network of digital systems working together, is important for enabling timely and reliable data flows. This is pertinent, for instance, in responding to crises, when access to accurate and up-to-date information is needed to inform responsive programming and decision-making. Without such digital infrastructure, data flows may be disrupted, or the data available may be inaccurate or incomplete.

Supporting countries on their AI journey

There is strong interest amongst Member States in adopting AI-powered technologies to improve people’s lives by providing better services. But as the benefits and risks of these technologies are uncovered, the need for an ethical data and AI governance framework, improved capacities and knowledge has become equally relevant. 

The ‘Joint Facility’ is an initiative launched by UNDP and ITU to enhance governments’ digital capacity development, including in harnessing AI responsibly. UNDP is assisting countries such as Kenya, Mauritania, Moldova and Senegal in developing data governance frameworks to promote the use of data for evidence-based decision making. Also under development is a ‘Data to Policy Navigator’ that is being created by UNDP and the BMZ's Data4Policy Initiative. The Navigator is designed to provide decision-makers with the knowledge they need to integrate new data sources into policy-development processes. No advanced or prior knowledge of data science is needed.

UNDP, along with UNESCO and ITU, is also part of a United Nations Inter-Agency Working Group on AI, where the goal is to share collective learnings and best practices for other countries’ benefit. The group has developed recommendations on AI Ethical Standards, which include key aspects of international and human rights regulations around the right to privacy, fairness and non-discrimination, and data responsibility. 

Countries are at different stages of their AI journey, and careful assessment is needed to determine the appropriate digital infrastructure, governance and enabling community that may be required based on their unique needs and capabilities. To this end, UNDP, along with Oxford Insights, designed an AI Readiness Assessment as a first step that can help countries better understand their current level of preparedness and what they may need moving forward as they seek to adopt responsible, ethical and sustainable AI systems.

The AI Readiness Assessment 

The AI Readiness Assessment comprises a comprehensive set of tools that allow governments to get an overview of the AI landscape and assess their level of AI readiness across various sectors. The framework is focused on the dual roles of governments as 1) facilitators of technological advancement and 2) users of AI in the public sector. Critically, this assessment also prioritizes ethical considerations surrounding AI use.

The assessment highlights key elements necessary for the development and implementation of ethical AI, including policies, infrastructure and skills. These aspects are important for countries to consider as AI-powered technologies are implemented at population scale to help meet national priorities and achieve the Sustainable Development Goals.