In recent years, the concept of Digital Public Infrastructure (DPI) has gained significant attention from the international community, including the UN and the G20, as a new policy paradigm for development. But understanding the risks of DPI is crucial to ensuring that its potential benefits materialise.The risks stem from the fact that “digital public infrastructure” lacks a clear definition. The term encompasses the many digital technologies that serve as economic and social infrastructure, from digital identification and payment systems to data exchanges and health services. As a policy initiative, though, DPI refers to a vague vision of using these technologies to serve the public interest. This could result in the Internet and technological innovation working for everyone – or just as easily turn them into tools for political control.In discussions about DPI, policymakers often point to cases that highlight how technology and connectivity can spur development. They frequently cite India’s Unified Payments Interface, which has expanded financial inclusion and reduced the costs of digital transactions for its hundreds of millions of users.It is also understood that such infrastructure is to be built with Digital Public Goods (DPGs), a concept that encompasses open-source software, open standards, and other non-proprietary components. This definition is partly intended to position DPIs as being “for the public” but also to enhance competition and mitigate concentrations of power in the global digital economy.Lastly, proponents point out that DPI could bolster international co-operation, particularly as the 20-year review of the World Summit on the Information Society (WSIS) approaches. This important UN initiative has provided the framework for countries to collaborate on digital development. Although authoritarian states have previously sought to assert greater control over the Internet’s governance during these negotiations, a focus on promoting DPI could avoid this politicised debate and instead foster a constructive agenda to bridge digital divides.But basing policy on such an ill-defined concept poses significant risks. Ideally, governments would convene other stakeholders to create an enabling environment for DPI and safeguard users’ rights and interests. It is easy to imagine, however, that some governments will place their own interests above civil liberties and fundamental rights, using this infrastructure for surveillance and targeting in the name of law enforcement or national security. An especially pernicious example could involve the monitoring and regulation of individual behaviour through dystopian social-credit systems.Moreover, while many proponents hope that DPI could chip away at Big Tech’s outsize power, it has also been associated with narratives of digital sovereignty that could contribute to the Internet’s fragmentation – a systemic threat to global communications. For example, one can imagine scenarios in which some governments challenge the multi-stakeholder model for governing global Internet resources like IP addresses and domain names on the grounds that they constitute DPIs.The policy vision of DPI will continue to evolve, and ongoing discussions, it is hoped, will help identify and clarify further opportunities and risks. Initiatives such as the UN’s Universal DPI Safeguards Framework, which seeks to establish guardrails for DPI, are a promising start. But much more must be done. For example, the UN’s framework has recognised the need for continuous learning to ensure that the right safeguards are in place.As the concept of DPI gains traction in the UN system and other multilateral organisations, vigorous and informed debate regarding its potential advantages – and pitfalls – will be essential. With clear-cut policy guidelines and protections, we can help prevent these technologies from becoming tools for surveillance and repression, ensure that everyone benefits from the burgeoning digital economy, and keep the Internet open, globally connected, and secure. – Project Syndicate