Heute 95

Gestern 137

Insgesamt 39278530

Freitag, 1.07.2022
eGovernment Forschung seit 2001 | eGovernment Research since 2001

When the concept of Smart City came into our lives, the technology used and the expected results were mainly in the direction of ensuring the interaction between the local government and the citizens to be served online. In addition, the provision of municipal services via online platforms or mobile applications and applications based on automation was considered smart city applications. At that time, how the big data produced by these applications should be used and its security were seen as one of the most important problem areas. Another was the inequality of access, which we describe as the 'digital gap', which was being seen based on various reasons. Although the issue of the digital gap is still valid, it has taken on a whole new meaning, as will be discussed below.

Today, the definition and content of the concept of Smart City have been transformed depending on the developing and widespread technologies. We no longer think of the Smart City as just piecemeal digitization of service delivery. Smart Cities are now mostly thought of as settlements where digital technologies and data are used for the purposes of economic growth, quality of life, and sustainability (Mora et al., 2019).

Moreover, new Smart City applications seem to focus on problem solving and resilience in the medium and long term. It is possible to examine the reasons for this situation in two directions or as the intersection of two different dynamics.

First, advancing technology offers us new possibilities every day. At the forefront of these possibilities is data processing with algorithms and machine learning, which leads us to artificial intelligence-assisted forecasting and decision making. Therefore, we can now benefit much more from large and detailed data sets.

Secondly, there is an increasing role and responsibility of cities in solving common problems of humanity and resilience against possible risks, which is a much more comprehensive and complex issue. As a result of globalization and the transformation of nation-state borders, cities at regional and global levels, especially global cities, have become the main scale for economic growth and for the solution of social, environmental, and spatial problems. In this context, cities choose the way to make decisions about the city by processing data more and better. On the other hand, as the data used in urban processes grows and diversifies, more and more urban areas are affected by artificial intelligence-based decisions.

Artificial Intelligence and Algorithm Bias

All these developments and the spread of artificial intelligence-based smart city applications bring new problems and difficulties too. In the most general terms, artificial intelligence algorithms used in urban processes also bring social responsibility (Falco, 2019). The widespread belief that technological advances and applications based on they are completely neutral in political, social, and economic terms cause smart city applications to be perceived as one-dimensional (Löfgren & Webster, 2020). Therefore, social and political factors that are ignored while focusing on technological solutions cause artificial intelligence-based decision-making processes to not achieve the targeted success. As solutions based solely on generating algorithms ignore the multi-factor structure of urban life and problems, one of the problems that arise is the bias that arises separately in algorithm logic and data sets or develops in relation to each other through training data (algorithmic bias). In his study, which focuses on the root causes of bias in algorithms and its reflections on smart cities, Gregory Falco states that the biased nature of algorithms stems from the data collection methods used, as well as the design of algorithms with the assumptions of programmers, makes AI biased (Falco, 2019).

Artificial intelligence experts have noticed the problem of algorithms giving uncalculated biased results for a while and have started to develop solutions. We know that the algorithms that feed the decision support systems can cause bias in the outputs in the machine learning process. Although this problem can be fixed by working on algorithms on the technical side, it is not possible to completely overcome this difficulty without solving the problem of data bias. Particularly, the bias of the training data used in the training of artificial intelligence is a problem beyond technology. For this reason, it needs to be resolved with different measures and methods. In particular, data sets consisting of feedback from city residents or complaints and similar inputs cannot provide the correct reflection of the actual problem to the data and thus to the algorithm. For this reason, decision support systems work in a biased way.

Another challenge is the bias caused by the digital gap. Since the day information and communication technologies entered our daily lives, the most important challenge we have faced on a social scale is the digital gap. The problem of the digital gap is actually a three-dimensional issue, namely access, use, and knowledge/skills, and this difference arises not only between individuals and households but also between business lines and geographical regions with different socioeconomic levels. We know that the COVID-19 pandemic has deepened and widened these gaps. Today, a new one has been added to these digital gaps.

The unequal distribution of data resulting from the unequal distribution in the use of smart city applications by individuals causes bias in decisions made based on artificial intelligence. This means a new and increased digital gap. The problem is no longer limited to just accessing online services.

Solutions to Challenges

In our age, while searching for ways to solve problems in many areas of daily life through technology, the emphasis on the real benefit of these technologies for everyone is becoming more evident each day. For everyone and for the benefit of all, the first step of technology is accessibility and the widespread use that comes with it. Especially today, data-based applications make this condition even more important. Although it is possible to find a solution to algorithmic bias, especially by controlling the data used in training algorithms and arranging tests, the most important steps to prevent bias in real-time data or constantly updated datasets are related to the dissemination of use to the society. Regarding the social dimensions of these difficulties, increasing the knowledge and usage skills of the society on artificial intelligence and data sharing will both increase the data quality and the correct decision-making rate of the algorithm, and will serve to form a participatory and democratic urban life, which is one of the main goals of Smart Cities. For this reason, the main areas are the ease of use of the applications, the introduction of the applications to the disadvantaged segments of society, especially in terms of digitalization, and the elimination of the concerns of the citizens about the protection of personal data or data security.

  • Falco, Gregory. "Participatory AI: Reducing AI bias and developing socially responsible AI in smart cities." 2019 IEEE International Conference on Computational Science and Engineering (CSE) and IEEE International Conference on Embedded and Ubiquitous Computing (EUC). IEEE, 2019.
  • Löfgren, Karl, and C. William R. Webster. "The value of Big Data in government: The case of ‘smart cities’." Big Data & Society 7.1 (2020): 2053951720912775.
  • Mora, L.; Deakin, M.; Reid, A. Strategic principles for smart city development: A multiple case study analysis of European best practices. Technol. Forecast. Soc. Chang. 2019, 142, 70–97.


Autor(en)/Author(s): Aysu Kes Erkul

Quelle/Source: Linked in, 17.05.2022

Bitte besuchen Sie/Please visit:

Zum Seitenanfang