Cities are sizing up the generative AI skyline. But first, they have to establish the ground rules

The use of generative artificial intelligence (GenAI) technology – which involves using data, algorithms and machine learning to generate new content – is surging among individuals and businesses; governments are also entering the arena by integrating GenAI into their operations. The GenAI market is projected to expand to $255.8 billion by 2033, compared to $13.5 billion in 2023.

As cities worldwide recognize the technology’s potential, several of them are formulating policies to regulate its use. Their development represents a crucial step in navigating the opportunities and challenges presented by emerging technologies. From Dubai to California’s San Jose, cities are exploring AI and GenAI governance with a focus on ethical considerations, transparency and accountability. These overarching frameworks represent the initial phase in an ongoing effort to develop and refine GenAI policies and procedures, aimed at fostering responsible use of GenAI capabilities in public administration and services.

Cities seem to be in the exploratory phase of adopting GenAI. In 2023, Bloomberg Philanthropies surveyed 100 mayors and city staff to gauge their views on the technology. City leaders worldwide expressed interest in utilizing it, with 69% of cities either exploring or testing GenAI capabilities.

These city officials are particularly interested in understanding how GenAI can help enhance citizen engagement, advance data-driven policy-making, optimize services and resource allocation, and streamline administrative processes and communication. In addition, they are interested in determining how GenAI can enhance city services and efficiency, especially in areas such as traffic and transportation, infrastructure, public safety, environmental issues, education and administrative tasks.

However, there seems to be a gap between interest and implementation. While 96% of mayors are interested in the use of GenAI, many have not developed comprehensive policies related to it, and only 2% of cities are actively adopting the technology. Those cities adopting GenAI are designating GenAI leads, creating high-level guidelines and providing training for their staff.

An initial preliminary review of GenAI city policies by the G20 Global Smart Cities Alliance in early 2024 indicates that some cities, primarily from North America and Europe, are taking early steps to develop policies and guidelines for governance of the technology.

Common principles among these guidelines include the prioritization of ethical considerations such as security, transparency, privacy and accountability. The guidelines emphasize the need for transparency in decision-making processes, responsible data use, and protection of individual rights. Moreover, there is a shared recognition of GenAI's potential impact on citizens and the importance of maintaining fairness, inclusivity and human oversight in the deployment of these technologies.

To share a few examples:

• Seattle, Washington has a policy that aims to establish requirements that city departments must follow when procuring and utilizing software that falls under the definition of GenAI.

• San Jose, California has a guideline setting baseline rules for users (such as staff, contractors and volunteers) while using GenAI for city work. In addition to usage of the technology being subject to the Public Record Act, the measure sets out principles for using GenAI, addresses copyright and content ownership, and requires users to cite AI when used and to understand the related risks.

• With concerns over the impact and use of GenAI, Boston, Massachusetts has an interim policy in place as a resource for employees when using such tools.

• Cities like Amsterdam and Helsinki have established registries to outline how algorithms are used in delivering services, while New York City has established laws requiring companies to disclose algorithms related to hiring.

The governance of GenAI by local government is relatively new – many of the ethical and responsible principles and practices being explored align with fundamental concerns related to technology governance.

Establishing responsible GenAI governance policies will be crucial in developing trusted products and services that can deliver the greatest value to society. The World Economic Forum’s AI Governance Alliance continues to lead on AI governance by fostering consensus on global best practices and ensuring technology governance contributes positively to societal well-being. Stakeholders can also refer to the AI procurement in the box tool to rethink how they acquire AI and make responsible purchasing decisions.

The exploration and adoption of GenAI by cities is still in its infancy, and its impact remains to be fully understood. Cities globally are facing numerous social, environmental and economic challenges. Technology, when used appropriately and when it is found to be the right and best available solution, can be a powerful tool for addressing them.

The potential of combining unbiased and large datasets with machine learning and other technological capabilities can bring significant benefits to society, governments and businesses. However, this potential can only be realized when risks and unintended consequences, both short- and long-term, are accounted for and addressed through governance frameworks designed to minimize harm. Additionally, finding cost-effective and practical use cases and learnings for cities is essential for successful adoption.

As shown by Gartner, it seems that GenAI is at the peak of the hype cycle and will take years before reaching the “plateau of productivity” where real adoption is likely to be observed. This presents opportunities to understand the technology’s potential in local government, thoroughly determining the barriers and risks early on, knowing when GenAI is and isn’t appropriate for a city challenge, and thinking about appropriate and agile solutions that can be sustainable over time and given complexities.

Cities are already encountering several obstacles to determining the true potential of GenAI, including knowing where it can make a difference. Mayors and city staff are also identifying barriers to adoption of the technology, including a lack of technical expertise, policies and awareness, budgetary constraints, data privacy and security concerns, limited access to sufficient infrastructure, and a lack of use cases. These challenges underscore the necessity for comprehensive strategies to tackle the complexities and risks associated with integrating GenAI into city operations and services.

Looking ahead, the future of GenAI policies in cities lies in the establishment of comprehensive frameworks that balance innovation with ethical considerations. As cities continue to navigate the complexities of GenAI governance, it is crucial for policy-makers to engage in ongoing dialogue with experts, stakeholders and the public to ensure that GenAI policies reflect the values and needs of the communities they serve.

Cities must prioritize the development of GenAI policies that address not only the technical aspects of its implementation, but also the ethical, legal and social ramifications of deploying such technology in an urban context. This involves fostering a culture of responsible use of GenAI, ensuring that systems are transparent, accountable and aligned with societal values. Cities should also focus on building technical expertise within their administrations and fostering partnerships with academia, industry and civil society.