Governing with artificial intelligence: Are governments ready?

Governments worldwide are exploring how artificial intelligence (AI) can transform public governance and improve government services. The OECD’s policy paper, “Governing with Artificial Intelligence: Are Governments Ready?” assesses the opportunities and challenges associated with integrating AI in the public sector. It outlines the significant benefits AI can bring alongside the risks and policy issues that must be managed. Here, we summarise the paper’s main findings and policy recommendations.

REPORT: Governing with artificial intelligence: Are governments ready?

The potential of AI in public governance

There are three main potential transformative benefits of AI in the public sector:

1. Productivity enhancement: AI can streamline internal government operations and improve the effectiveness of public policies. It can automate repetitive administrative tasks, allowing public officials to focus on more complex and strategic functions. For example, the Queensland Government in Australia uses AI to map and classify land use features from satellite imagery. This helps improve the response to biosecurity and natural disaster events and allows the Government to map and categorise land use more quickly and efficiently.

2. Responsiveness and inclusivity: AI can help governments design and deliver more proactive and personalised public services, making them more responsive to citizens’ needs. During the COVID-19 pandemic, Norway’s Labour and Welfare Administration significantly improved service delivery by using a conversational AI named Frida to assist citizens in accessing social benefits around the clock.

3. Accountability and oversight: AI can bolster government accountability by enhancing the capacity for oversight and red-flagging potential fraud risks. For example, AI can help detect fraudulent activities in public procurement and monitor the distribution of social benefits, thereby ensuring greater integrity and transparency. In Brazil, the federal government’s internal auditor developed Alice, a tool that automates procedures and enables continuous and large-scale auditing of public tenders.

Challenges and risks

At the same time, there are several risks associated with the responsible deployment of AI in the public sector:

1. Bias and fairness: If not carefully managed, AI systems can perpetuate — or even amplify — existing biases. This can lead to unfair and discriminatory outcomes, particularly in sensitive areas such as law enforcement and welfare benefits. Canada requires the publication of completed Algorithmic Impact Assessments to enhance transparency and accountability. In the Netherlands, the Algorithm Register has similar objectives.

2. Transparency and explainability: Ensuring that AI decisions are transparent and explainable is crucial for maintaining public trust. The complexity of AI systems often makes it difficult to understand how decisions are made, posing a challenge to accountability and traceability of responsibilities. Many governments, national and subnational, are thus putting in place open registers of public algorithms, such as in Chile and France.

3. Data privacy and security: AI involves processing vast amounts of data, raising significant concerns about data privacy and security. Governments must ensure robust data governance frameworks to protect citizens’ privacy and personal information. The quality and interoperability of data in the public sector are major challenges for many governments, prompting many to closely link AI and data strategies.

Align strategies with public sector goals, build capacities and set standards

To address these risks and challenges, the OECD outlines several policy recommendations for governments:

1. Strategic frameworks: Governments should develop comprehensive AI strategies aligned with broader public sector goals. These strategies should include clear objectives, central steering, and continuous monitoring and evaluation mechanisms.

3. Regulatory standards: Robust regulatory frameworks and ethical guidelines, including standards for algorithmic transparency, data privacy, and bias mitigation, are crucial for ensuring AI’s safe and responsible use.

2. Capacity building: Enhancing the digital skills of public sector employees is essential for effective and responsible AI deployment and use. Training programmes and partnerships with academic institutions and the private sector can help build expertise and resource teams efficiently.

4. International collaboration: Given the global nature of AI challenges, international cooperation is vital. Governments should participate in global forums and initiatives to share best practices and develop common standards.

A strategic and responsible approach to AI in government

Integrating AI in the public sector promises to enhance productivity, responsiveness, and accountability. However, realising this potential requires a strategic and responsible approach that addresses the associated risks. The OECD’s policy paper provides a valuable roadmap for governments to navigate the complex landscape of AI governance, ensuring that AI is used in the public interest to benefit all citizens while safeguarding their rights.