AI for municipalities

Improving citizen services, speeding up routine tasks, and freeing up human resources for other tasks – this is what local authorities hope to achieve with artificial intelligence in administration. The topic is an important one and interest was high: The event recorded over 100 registrations, with nearly 60 participants joining online – from cities such as Lusaka (Zambia), Stuttgart, Mombasa (Kenya), Berlin, and Iserlohn.

Prof. Dr.-Ing. Mahdi Bohlouli from the AI company Petanux in Bonn explained the basics:

At its heart, AI’s main strength lies in its ability to analyse information, recognise patterns, create content on that basis, support decision-making, and even carry out tasks independently, he explained. In public administration, AI can thus relieve administrative processes (e.g. by automating routine tasks), facilitate urban planning (e.g. through AI-supported data analysis), improve citizen services (e.g. chatbots), and enhance cybersecurity (e.g. by detecting anomalies). However, this requires strong data protection and data sovereignty, ensured through local data storage and locally operated AI applications. Open-source products provide local authorities with cost-effective technical solutions.

municipalities must also consider ethical aspects such as transparency and accountability in AI decision-making, as well as the concerns and training needs of their staff. Clear internal guidelines, a transparent and participatory approach, cooperation between municipalities, and small pilot projects that deliver visible results help ensure a smooth and safe introduction of AI. For a start he suggested using AI for meeting minutes and AI-based knowledge management. Automating such and similar tasks can free up considerable human resources, which can be used to ease staff shortages elsewhere.

“Let AI handle the routine tasks, so you can concentrate on the complex ones,” he advised.

The example of Jakarta – presented by Andhika Ajie from the Regional Research and Innovation Centre in Jakarta – demonstrated how AI is being used to tackle the challenges of a transforming megacity.

The long-term plan, ‘Jakarta Spatial Plan Development,’ considers digitally driven development as one of three key pillars for municipal development. The corresponding framework, ‘Jakarta Smart City’ places innovation and citizen well-being at the top of its priorities. “Technology is a tool. It should enhance urban life and focus on the needs of people, “ says Andhika Ajie. Jakarta therefore pursues a collaborative process in which all city stakeholders – citizens, the media, academia, businesses, and authorities – work together.

The city has developed a so-called super app – Jakarta Kini, or JAKI for short – Andhika Ajie reported. This digital ecosystem integrates a wide variety of applications and provides residents with a one-stop platform for requesting services (such as health and social services), accessing information (such as official announcements, food prices, public transport) and reporting issues (such as infrastructure deficiencies).

Also, given Jakarta’s enormous population density, mobility and traffic management is a key area for AI applications. The city is using a CCTV monitoring service including 1,500 cameras across the city to measure crowd and vehicle density. The data serve as the basis for AI to regulate road traffic. In addition, the city is testing AI-driven real-time management of public transport through the pilot Passenger Load Intelligence System (PLIS).

Another important application focusses on flood management – urgently needed in Jakarta, which is rapidly sinking below sea level. A machine-learning-based system (a system that learns from examples and enables predictions) – analyses data on water levels and flooded areas to coordinate protective measures and emergency response efforts.

The example of Hamburg – presented by Lisa Eglhofer from the Sustainability Unit in Hamburg finally delved into the details of AI-supported development of a sustainability strategy.

In its new strategy paper, the city aims to build directly on the current implementation of the 17 Sustainable Development Goal, taking all existing concepts and activities into account. The database to be used is the city’s transparency portal, which however contains the impressive number of 170,000 documents. An AI analysis tool is now to assist with the time-consuming task of reviewing and processing the data. Technically, the system is based on various large language models (LLMs), such as Gemini 1.5 Pro and Flash, GPT-4, and GPT-5). For its training, the Sustainability Unit manually reviewed, grouped, and coded over 80 documents for their relevance to the SDGs.

The results are twofold: Through a dashboard, the AI now enables micro-level analysis of individual documents in terms of their links to the SDGs. In addition, the analysis tool generates fact sheets on the progress of individual SDGs or their sub-goals. The underlying process works as follows: An AI Agent Planner coordinates the research request. An AI Agent Researcher gathers the relevant data in three iterative loops. And an AI Agent Synthesizer compiles the results. These fact sheets then will serve as the basis for developing the sustainability strategy.

For the technical side, Hamburg secured funding and brought in the support of an AI agency. Thus, the project turned out to be not that difficult after all. In the future, the city intends to offer the code as an open-source product to other municipalities. The plan is also to expand the dataset, to connect the tool to budget planning data, and to make it even easier to use.

„In my experience, it’s often difficult to secure funding for sustainability. Now, that we talk about AI and digital approaches, it’s much easier to get politicians on our side.“ (Lisa Eglhofer)