As artificial intelligence grows more ubiquitous, its potential and the challenges it presents are coming increasingly into focus. How we balance the risks and opportunities is shaping up as one of the defining questions of our era. In much the same way that cities have emerged as hubs of innovation in culture, politics, and commerce, so they are defining the frontiers of AI governance.
Some examples of how cities have been taking the lead include the Cities Coalition for Digital Rights, the Montreal Declaration for Responsible AI, and the Open Dialogue on AI Ethics. Others can be found in San Francisco’s ban of facial-recognition technology, and New York City’s push for regulating the sale of automated hiring systems and creation of an algorithms management and policy officer. Urban institutes, universities and other educational centres have also been forging ahead with a range of AI ethics initiatives.
These efforts point to an emerging paradigm that has been referred to as AI Localism. It’s a part of a larger phenomenon often called New Localism, which involves cities taking the lead in regulation and policymaking to develop context-specific approaches to a variety of problems and challenges. We have also seen an increased uptake of city-centric approaches within international law frameworks.
In so doing, municipal authorities are filling gaps left by insufficient state, national or global governance frameworks related to AI and other complex issues. Recent years, for example, have seen the emergence of “broadband localism”, in which local governments address the digital divide; and “privacy localism”, both in response to challenges posed by the increased use of data for law enforcement or recruitment.
AI localism encompasses a wide variety of issues, stakeholders, and contexts. In addition to bans on AI-powered facial recognition, local governments and institutions are looking at procurement rules pertaining to AI use by public entities, public registries of local governments’ AI systems, and public education programs on AI. But even as initiatives and case studies multiply, we still lack a systematic method to assess their effectiveness – or even the very need for them. This limits policymakers’ ability to develop appropriate regulation and more generally stunts the growth of the field.
Building an AI Localism framework
Below are ten principles to help systematise our approach to AI Localism. Considered together, they add up to an incipient framework for implementing and assessing initiatives around the world:
Principles provide a North Star for governance: Establishing and articulating a clear set of guiding principles is an essential starting point. For example, the Emerging Technology Charter for London, launched by the mayoral office in 2021 to outline “practical and ethical guidelines” for research around emerging technology and smart-city technology pilots, is one example. Similar projects exist in Nantes, France, which rolled out a data charter to underscore the local government’s commitment to data sovereignty, protection, transparency, and innovation. Such efforts help interested parties chart a course that effectively balances the potential and challenges posed by AI while affirming a commitment to openness and transparency on data use for the public.
Public engagement provides a social license: Establishing trust is essential to fostering responsible use of technology as well as broader acceptance and uptake by the public. Forms of public engagement – crowdsourcing, awareness campaigns, mini-assemblies, and more – can help to build trust, and should be part of a deliberative process undertaken by policymakers. For example, the California Department of Fair Employment and Housing held their first virtual public hearing with citizens and worker advocacy groups on the growing use of AI in hiring and human resources, and the potential for technological bias in procurement.
AI literacy enables meaningful engagement: The goal of AI literacy is to encourage familiarity with the technology itself as well as with associated ethical, political, economic and cultural issues. For example, the Montreal AI Ethics Institute, a non-profit focused on advancing AI literacy, provides free, timely, and digestible information about AI and AI-related happenings from across the world.
Tap into local expertise: Policymakers should tap into cities’ AI expertise by establishing or supporting research centres. Two examples are the Confederation of Laboratories for Artificial Intelligence Research in Europe (CLAIRE), a pan-European project that takes a European focus to AI uses in cities and “How Busy Is Toon”, a website developed by Newcastle City Council and Newcastle University to provide real-time transit information about the city centre.
Innovate in how transparency is provided: To build trust and foster engagement, AI Localism should encompass time-tested transparency principles and practices. For example, Amsterdam and Helsinki disclose AI use and explain how algorithms are employed for specific purposes. In addition, AI Localism can innovate in how transparency is provided, instilling awareness and systems to identify and overcome “AI blind spots” and other forms of unconscious bias.
Establish means for accountability and oversight: One of the signal features of AI Localism is a recognition of the need for accountability and oversight to ensure that principles of responsive governance are being adhered to. Examples include New York City’s Algorithms Management and Policy Officer, Singapore’s Advisory Council on the Ethical Use of AI and Data, and Seattle’s Surveillance Advisory Working Group.
Signal boundaries through binding laws and policies: Principles are only as good as they are implemented or enforced. Ratifying legislation, such as New York City’s Biometrics Privacy Law, which requires clear notices that biometric data is being collected by businesses, limits how businesses can use such data. It also prohibits selling and profiting from the data sends a clear message to consumers that their data rights and protections are upheld and holds corporations accountable to respecting privacy privileges.
Use procurement to shape responsible AI markets: As municipal and other governments have done in other areas of public life, cities should use procurement policies to encourage responsible AI initiatives. For instance, the Berkeley, California Council passed an ordinance requiring that public departments justify the use of new surveillance technologies and that the benefits of these tools outweigh the harms prior to procurement.
Establish data collaboratives to tackle asymmetries: Data collaboratives are an emerging form of intersectoral partnership, in which private data is reused and deployed toward the public good. In addition to yielding new insights and innovations, such partnerships can also be powerful tools for breaking down the data asymmetries that both underlie and drive so many wider socio-economic inequalities. Encouraging data collaboratives, by identifying possible partnerships and matching supply and demand, is thus an important component of AI Localism. Initial efforts include the Amsterdam Data Exchange, which allows for data to be securely shared to address local issues.
Make good governance strategic: Too many AI strategies don’t include governance and too many governance approaches are not strategic. It is thus imperative that cities have a clear vision on how they see data and AI being used to improve local wellbeing. Charting an AI strategy, as was undertaken by the Barcelona City Council in 2021, can create avenues to embed smart AI use across agencies and open up AI awareness to residents to make responsible data use and considerations a common thread rather than a siloed exercise within local government.
AI Localism is an emergent area, and both its practice and research remain in flux. The technology itself continues to change rapidly, offering something of a moving target for governance and regulation. Its state of flux highlights the need for the type of framework outlined above. Rather than playing catch-up, responding reactively to successive waves of technological innovation, policymakers can respond more consistently, and responsibly, from a principled bedrock that takes into account, the often competing needs of various stakeholders.