Techno-politics of using AI to manage flood risks in Jakarta

Octavianti, Thanti

Abstract

In a changing climate, cities are facing mounted pressure to cope with weather-related hazards that are getting more frequent and severe. Some cities have been incorporating advanced technology to inform decisions that are smart and timely to avert, mitigate and manage disaster risks. Drawing insights from Science and Technology Studies (STS), this study seeks to investigate how the digital infrastructures and data systems might improve as well as clash with the decades of managing flood risks conventionally in a post-colonial bureaucratic setting. Jakarta was selected as a case study because of its complex flood risks exacerbated by land subsidence, a recent investment in the smart flood control initiative (IDR 200 billion/GBP 11 million). The government uses AI primarily to better understand how flood events occur and hence enhance its early warning system. Data collection included key informant interviews and policy document analysis and focused on the early adoption of the new system.

We found that the introduction of the new system creates opportunities to divert from the path-dependent environment but the old system presents barriers to such a breakthrough. Whilst Jakarta has a smart city platform to facilitate digital transformation, institutional inertia is one of the main factors hindering this transformation. There has been a disconnection between how the city is envisioned (including the hype and expectation around AI adoption) and how stakeholders do their everyday job. For example, many officers still do a manual recap of river water levels as the sensor data are perceived to be inaccurate. Furthermore, the AI-powered system might create a new mechanism of power. Experts can, to a certain extent, hold central control acting at the top of a rigid social hierarchy, making the role of experts a political one.

• Presentation Conference Type: Conference Paper (unpublished)
• Conference Name: The RGS-IBG Annual International Conference
• Start Date: Aug 31, 2023
• Deposit Date: Sep 15, 2023
• Publicly Available Date: Sep 15, 2023
• Public URL: https://uwe-repository.worktribe.com/output/11076059