OECD launches G7 toolkit for ‘safe, secure and trustworthy’ AI in the public sector
The Organisation for Economic Co-operation and Development (OECD) has published a new policy document for member states of the G7 to use as a guide when applying artificial intelligence (AI) technologies across the public sector.
The OECD called the AI ‘toolkit’ “a comprehensive guide designed to help policymakers and public sector leaders translate principles for safe, secure and trustworthy AI into actionable policies”.
The paper builds on the creation of national AI strategies across G7 countries. Drawing on data gathered from a questionnaire, it compares AI use cases across member states to identify cross-cutting themes in AI deployment to offer best practice.
The document says that while governments play a key innovative role and are often “prime users of AI”, more needs to be done to address concerns around public trust and safety.
Governments are also urged to think about AI development with respect to how it could impact human rights, fundamental freedoms, and the rule of law.
Other ethical risks raised include the unintended consequences of “bias and discrimination” embedded in AI systems, as well as “denial of individual autonomy”, “invasion of privacy”, and “infringement of intellectual property rights”.
Read more: OECD urges countries to gear up for ‘governing with AI’
Effective governance
On the subject of AI governance, the paper finds that while various national AI strategies reveal similar challenges in deploying AI, G7 members have “different institutional arrangements to govern the development and use of AI in the public sector”.
“Half of the members adopt a decentralised approach…while the other half report having adopted a more centralised approach [with] a dedicated single leading institution, or assigned a lead or coordinating role to an existing national body or organisation often upgraded to the level of ministry,” it says.
It adds that these arrangements depend on countries’ “existing institutional context and culture”. This may involve “different entities with varied coordination mechanisms and responsibilities across leading institutions”.
Also highlighted is the tendency of G7 members to mitigate tensions over “mandates, structures and mechanisms” across siloed public sector organisations by creating multi-institutional approaches or “lead institutions” with coordination roles.
Examples of these governance structures include the US federal government’s two key councils created to oversee AI initiatives across federal agencies. These include an interagency council, whose role is to “coordinate the integration of AI into agency programmes and operations”, and an executive-level council made up of cabinet members or their appointees, whose job is to “coordinate agency activities throughout the federal government”.
Examples of single institutional approaches come from Germany and the UK. Germany created “a central contact and coordination point for AI projects and initiatives” through an Advisory Centre for Artificial Intelligence under its Ministry of Interior. The UK, meanwhile, unified its efforts to support digital transition of public services across government through the Department for Science, Innovation and Technology.
Read more: The UK civil service isn’t ready for AI yet – but there are reasons for optimism
Good data, good AI
Data quality is highlighted as a crucial component of an effective AI toolkit. The paper says that G7 countries have become increasingly conscious of this connection and are “strategically linking their AI and data strategies to ensure that AI systems are built on robust, accurate and comprehensive datasets, in particular government data, including open government data”.
Both open government data and investment in data centres are found to be common themes of G7 governments’ AI strategies. The paper explains the purpose of both as to “facilitate data sharing and interoperability between different government departments and public sector organisations”.
Canada and Germany are noted for prioritising the use of open government data in AI applications. For example, Germany’s Open Data Act, enacted in 2017, mandates federal authorities to “make their data openly accessible by default”. In Japan, the government applies what it calls Basic Principles on Open Data, which has since been reviewed to keep pace with AI.
Italy has done similar work. The country’s AI strategy outlines a plan to use targeted data relevant to “particular use cases” of AI, which the Italian government expects will produce AI applications capable of addressing “specific needs and challenges within public services”.
The US government’s Office of Management and Budget has advised agencies to develop “adequate infrastructure and capacity to sufficiently share, curate and govern agency data for use in training, testing and operating AI”.
Finally, the UK has employed several key data initiatives, which stress “open standards” to make data readily accessible and reusable. These standards cover metadata, various data formats, and application programming interfaces (APIs).
About Jack Aldane
Jack is a British journalist, cartoonist and podcaster. He graduated from Heythrop College London in 2009 with a BA in philosophy, before living and working in China for three years as a freelance reporter. After training in financial journalism at City University from 2013 to 2014, Jack worked at Bloomberg and Thomson Reuters before moving into editing magazines on global trade and development finance. Shortly after editing opinion writing for UnHerd, he joined the independent think tank ResPublica, where he led a media campaign to change the health and safety requirements around asbestos in UK public buildings. As host and producer of The Booking Club podcast – a conversation series featuring prominent authors and commentators at their favourite restaurants – Jack continues to engage today’s most distinguished thinkers on the biggest problems pertaining to ideology and power in the 21st century. He joined Global Government Forum as its Senior Staff Writer and Community Co-ordinator in 2021.
Related Posts
