An investigation into the application of artificial intelligence (AI) by local governments across the globe has revealed significant insights into both the uses and potential pitfalls of these technologies. A study encompassing 170 local councils has documented 262 instances of AI being deployed in public service delivery, highlighting its growing role in governance. However, with this expansion comes an alarming lack of transparency and oversight surrounding its implementation.
One notable case occurred in Rotterdam, Netherlands, where an AI system was used in 2017 to assess the likelihood of welfare recipients committing fraud. This system, unfortunately, developed biases, disproportionately flagging individuals who were female, young, had children, and had limited proficiency in the Dutch language as “high risk.” Following a thorough ethics review, the use of this AI system was suspended in 2021. This situation underscores the dangers inherent in the adoption of AI technologies without adequate regulatory frameworks.
The study identified five key domains in which AI is currently making an impact:
Administrative Services: In Madrid, Spain, the VisitMadridGPT tourism chatbot offers users personalised recommendations, providing real-time assistance and cultural insights.
Health Care and Wellbeing: The Boston mayor’s office responded to COVID-related challenges by introducing an AI-driven chatbot for contactless food delivery, addressing both hunger and safety concerns during the pandemic.
Transportation and Urban Planning: Logan City, Australia, has implemented a real-time AI system to inform drivers of available parking spots, thereby alleviating traffic congestion and frustration. Additionally, AI technologies for route optimisation in public transport are gaining traction to enhance efficiency and reduce emissions.
Environmental Management: In Hangzhou, China, an AI system has been introduced to improve waste classification, effectively increasing recycling rates.
Public Safety and Law Enforcement: In Chicago, USA, AI integrated with sensors has been utilised to inform law enforcement tactics, resulting in a reported 25% reduction in gun violence in 2018. However, this use of technology has not been without controversy due to concerns regarding potential racial profiling.
Despite the advances being made, a concerning trend has emerged: only 26 of the 170 governments studied—around 16%—had established published AI policies by May 2023. Most local governments reportedly deployed powerful AI systems without public oversight or accountability frameworks, which raises significant concerns regarding ethical practices, systemic biases, and responsible data usage.
The gap in public awareness regarding AI technologies employed by local governments is also noteworthy. A recent survey indicated that over 75% of respondents had some knowledge of AI's presence in everyday life, yet many were unaware that their local councils were actively utilising these systems. Approximately 68% reported no knowledge of local AI governing policies, indicating a pressing need for improved transparency and communication from local authorities.
Among local councils that have established AI policies, Barcelona City Council has distinguished itself by promoting transparency, explanation of AI decisions, and fairness, thus offering a possible benchmark for other municipalities.
As local governments worldwide progress in the integration of AI into urban management—ranging from traffic monitoring to public safety—the need for robust guiding principles becomes increasingly essential. The implications of these technologies hold the potential for significant improvements in governance. However, without an emphasis on accountability, transparency, and ethical usage, cities may risk exacerbating existing inequalities and perpetuating biases.
Efforts are underway as researchers work with local governments in countries such as Australia, Spain, the US, Hong Kong, and Saudi Arabia to formulate guiding principles for responsible AI integration, with a goal to establish these by the end of 2025. The study reflects a pressing need for local governance to develop comprehensive policies aimed at ensuring that the application of AI serves the public good while safeguarding against its inherent risks.
Source: Noah Wire Services