Adoption of artificial intelligence (AI) in public administrations is growing, but governments face a range of challenges in scaling up AI applications, leading to many initiatives remaining in their pilot phase, according to a new OECD report.
As AI becomes an integral part of digital government, transforming processes and services to improve productivity, responsiveness and accountability, the OECD's Governing with Artificial Intelligence report is the first of its kind to examine how governments use and engage with AI, analysing 200 use cases and dozens of policy initiatives to understand how AI is applied and promoted across 11 core areas of government.
AI has shown benefits in areas such as automation, with AI-powered chatbots being used to answer citizen queries and assist with form filling; forecasting, with AI helping to anticipate natural disasters and speed up response efforts; and improved decision making, with AI-powered platforms clustering public opinions and identifying consensus statements during consultations. However, governments must ensure AI is used in a trustworthy manner, promoting transparent and responsible use. They also face unique barriers that hinder the rapid uptake of AI, including skills shortages, difficulty accessing and sharing quality data, financial costs, outdated regulations, and older IT systems.
The use of AI in government varies across functions: the report finds that AI is most commonly used in public services, civic engagement, and justice functions. Adoption lags in areas such as civil service management and policy evaluation. Tax administration makes moderate use of AI, as legal and regulatory complexities, particularly rules on data governance, may slow the uptake of the most advanced generative AI systems.
Unlocking the various benefits of AI requires governments to balance innovation with responsibility. Risks exist in the form of skewed outcomes, technical failures, data protection non-compliance and the exclusion of certain groups - challenges that, if left unaddressed, could undermine the effectiveness of AI use in government and citizen trust in government. However, not adopting AI also carries risks, such as missed opportunities for further efficiency and effectiveness of government, and a widening gap between public and private sector capacity.
"With more and more governments incorporating AI into service delivery, the right policy frameworks and safeguards are essential to encourage innovation and ensure the trustworthy use of AI" said OECD Secretary-General Mathias Cormann. "The OECD's Framework for Trustworthy AI in Government provides guidance on the responsible development, deployment and use of AI by public sector entities."
The report identifies governance, data, digital infrastructure, skills, investment, procurement, and partnerships with non-government actors as the seven key enablers for the successful adoption of AI in government. A co-ordinated approach to AI transformation across government that is user-centred, engages meaningfully with stakeholders and the public, and establishes clear but proportionate guardrails will manage risks while remaining agile and adaptable to future changes.
Explore the OECD's other work on Artificial intelligence.
