Deploying artificial intelligence is attractive for municipalities
Research and development in the field of artificial intelligence (AI) have made a quantum leap in recent months. The focus has primarily been on ChatGPT (Generated Pre-trained Transformer). The version, which was released in November 2022, already caused furore among experts, and now there have been updates that provide helpful digital support in many areas of life. As a result, diverse tasks can now be performed by AI-supported tools: they provide text descriptions of images, program websites with shops, write speeches, research, develop policies, draw up presentations, create deceptively real-looking photos and videos, compose music and design logos – there is no limit to the imagination. AI is performing services that previously appeared to be solely the realm of human reasoning and human creativity. However, looking at it globally also revealed fake photos created by AI that quickly went viral on social media, e.g., Pope Francis in a trendy down jacket or fleeing from the police. It makes one of the biggest risks of AI clear: how can you tell the difference between an original and a fake, and what are the consequences if people believe fake news is the truth?
Currently, the AI developers and financiers themselves are also attracting attention: on the one hand, prominent experts from the tech elite, in particular from the USA, are calling for a development pause for AI tools. There is uncertainty and concern regarding possible negative and irreversible consequences of the exponentially developing systems. On the other hand, leading AI developers are already positioning themselves in the global emerging market of artificial intelligence. The line between matters of ethical responsibility and uncompromising economic interests is becoming increasingly blurred.
Artificial intelligence – Key and risk technology simultaneously
It is not just the dynamic of the technological-economic development that is astonishing but also the change in the debate around AI. Until just a few months ago, expert papers on the topic claimed that copying or even exceeding human intelligence by way of a “strong AI” was most likely a hypothetical future scenario. At the moment, the question is being asked whether further development of AI systems is actually still responsible while we are unable to imagine with any certainty that its effect on humanity is positive and the risks are controllable.
The only thing that appears clear is that AI harbours both great potential and extraordinary dangers, although as with every technology the biggest danger is humanity itself. Therefore, the debate on values that has flared up is extremely urgent, just as the German Ethics Council made clear in its most recent opinion. Existing initiatives, such as the German Federal Government’s AI strategy or the EU Commission’s planned single legal framework for the use of AI, have to stay abreast of the rapid developments.
Artificial intelligence applications are being increasingly deployed in the municipalities
Artificial intelligence stands for technologies and applications that reproduce essentially human traits: the ability to observe, to analyse and to make decisions based on reason. The key to it is “machine learning”, which on the basis of mass data (big data) combined with automated statistical methods is capable of recognising patterns, mapping models and learning. A branch of machine learning is “deep learning” in which artificial neuronal networks learn with the help of large amounts of data.
Away from the current debate around strong AI, which is close to or even exceeds human intelligence, applications using “weak AI” have long since become part of our daily lives and are taken for granted, e.g., image and voice recognition, navigation and digital assistant systems. That can also be seen in the municipalities where the deployment of artificial intelligence and self-learning algorithms based on complex amounts of data in real time are to be shown in the term smart city. For example, there may be gains in efficiency in administration thanks to the improved analysis of data, documents and processes. Interactions with citizens could be made more service-orientated through AI-supported public authority platforms, online citizen services or dialogue assistant services. E-participation platforms and community apps to forms of augmented and virtual reality expand the spectrum of participation and communication formats, for example by being able to address different target groups according to their user profiles. Urban development planning and urban planning departments can gain a better decision-making basis thanks to automated trawling and merging of databases as well as pattern recognition in large amounts of data. Citizen’s requests could be better considered with an AI-supported structuring of statements.
Digital twinning (the virtual representation of physical objects online) could be a source of inspiration for creative decision processes as it could be used to depict planning projects and necessary process workflows. In the field of public safety, there are the possibilities arising from sensor-based control and monitoring. AI’s current deployment in the field of mobility is of particular importance regarding condition evaluation of roads or the traffic flow analysis and forecast. In the future, automated public transport or even intelligent transport systems will be of greater importance. In the energy sector, it is primarily a question of the better control of bi-directional energy flows including the necessary storage systems and also the communication between public energy grids and building services technology.
Useful AI requires testing and evaluation
In order to ascertain how the deployment of AI systems is useful on a case-by-case basis, it must be considered against the backdrop of the local conditions. Many applications may not prove to be useful during testing, while other applications might demonstrably improve the quality of life in our cities. Challenges are found, for example in regard to potential discriminating effects in the selection of algorithms and the connected transparency of decision and optimization processes as well as in the fields of monitoring and evaluation. Automation requires particular care also in regard to the operation of critical infrastructures (keyword: cyber security).
The question of suitable political or legislative frameworks can therefore not be answered alone on the international, European and national levels. It must be intensively addressed locally, precisely where AI can be realistically experienced. Digitalisation is generally not just a question of the technically possible, but above all is also a question of the individual, organisational and societal attitude and mindset regarding the deployment of digital tools, and the same applies for the deployment of AI. For the municipal administrations and the provision of public utilities and services, the multi-faceted application opportunities offer a considerable potential to ease the burden. When time-consuming routine tasks can be increasingly accomplished by computers, it creates freedom in the deployment of employees. At the same time, it is to be expected that the problem solving abilities of specialist departments, such as urban planning, will clearly increase through the process of machine learning.
Early consideration of the possible risks of AI deployment
For the successful deployment of AI in municipalities, a series of basic conditions are necessary: opening of digital databases in machine-readable formats, generally regarding the municipal companies as a digital ecosystem, a lived municipal data governance (including data ethical guidelines and open data), establishing data platforms with standards and interfaces, expanding data and AI competences as well as generally a culture of openness towards the digital administration with the deployment of AI. By necessity, it includes a critical reflection on the risks and challenges accompanying the use of AI. It requires a dialogue so that the job descriptions in the administration and also in the municipal enterprises change. Therefore, the need to be retrained or further educated is to be recognised in good time. In relation to urban planning, it must focus on better understanding the impact of increasingly more sensors and algorithmically controlled municipal processes. However, in terms of policy-making the question is asked, how increasingly more available real-time information and the growing danger from the blurred distinction between knowledge and fake information influence the work in the municipal councils and expert committees.