From Trustworthy AI Principles to Public Procurement Practices

This book is an early warning to public officials, policymakers, and procurement practitioners on the impact of AI on the public sector. Many governments have established national AI strategies and set ambitious goals to incorporate AI into the public infrastructure, while lacking AI-specific procurement guidelines. AI is not traditional software, and traditional processes are not sufficient to meet the challenges AI brings. Todays decisions to embed AI and algorithmic systems into public system infrastructure can and will have serious repercussions in the future.
The promise of AI systems is to make the public sector more efficient, effective, fair, and sustainable.
However, AI systems also bring new and emerging risks which can impact rights and freedoms. Therefore, guardrails are necessary to consider the socio-technical dimensions and impact on individuals, communities, and society at large.
It is crucial that public sector decision-makers understand the emerging risks of AI systems, the impact on the agency and the wider public infrastructure, and have the means to independently validate vendor claims. This book is a result of interviews with more than 20 public procurement professionals across countries, offering an in-depth analysis of the risks, incidents, governance practices, and emerging good practices around the world, and provides valuable procurement policy and process recommendations to address and mitigate these risks.