Algorithmic City? Acceptance of Algorithms in the Context of Urban Governance

This paper aims to understand and empirically measure the acceptance of algorithmic urban governance among residents of smart cities. Using data from representative surveys in Singapore, Tallinn and Warsaw, we propose the composite indicator ‘AI acceptance index’ (AIAI) to illustrate differences between cities and socio-economic groups. We found significantly higher acceptance of AI implementation among Singaporeans than Varsovians. In Warsaw, respondents were much less accepting to AI solutions in welfare system, public investments as well as in making decisions about which demonstrations should not take place due to the potential risk of riots. However, they were ready to accept such solutions in making decisions about the public transport network or schedule or where more police patrols are sent. Meanwhile, in Singapore, people were more eager to accept the implementation of AI in making welfare decisions, directing public investments as well as deciding which demonstrations should be allowed. The analysis also indicates patterns reflecting social differences in AI acceptance. In both Warsaw and Singapore, people over 45 years were declaring lower AI-acceptance, while highest AIAI was observed among people under 35 years. In Singapore, respondents with college education declared visibly higher AIAI than others. The strong similarity between the cities was observed in the case of financial situation – better-off respondents were also more positively inclined towards AI implementation in urban decision-making. Those preliminary results suggest that attitudes towards smart technologies, like AI algorithms, are shaped by the socio-economic status, but not in a straightforward way. For example, in Warsaw, contrary to Singapore, people with lower education were more accepting to AI. In both cities, the youngest cohorts were more techno-optimistic and open to the development of algorithmic urban governance. After testing and piloting our analytical approach, we believe that our results signal general patterns present in smart cities worldwide.