Page 8 - Unlocking the potential of trust-based AI for city science and smarter cities

denial-of-service attacks, disrupting the sensing, transmission, and control to degrade the quality
of intelligent services in a smart city. In addition, the pervasive video surveillance in a smart city
captures a tremendous number of images and video clips, which may be utilized to infer local
residents’ trajectories and inherently endanger their privacy.

The home area information collected and managed by smart home applications may pave the way
to disclosing residences’ highly privacy sensitive lifestyle and even cause economic loss. Although
some off-the-shelf techniques (encryption, authentication, anonymity, etc.) and policies might be
directly applied to avert these problems [5], the emerging “smart” attackers could still infer and
violate privacy in many other ways, such as side channel attack and cold boot attack [6]. Without
sufficient security and privacy protections, users may refrain from accepting the smart city, which
would remain as a far-off futuristic idea.

Figure 1. Smart city applications [7].

1.2 Challenge and response

Smart cities provide services that benefit from the city-scale deployment of sensors, actuators, and
smart objects. Such services are mainly driven by data and can be broadly classified as producers
of data, consumers of data, or a combination of both. For example, a parking service that deploys
a message queue telemetry transport (MQTT) broker to publish parking lots’ availability data is
considered a producer, while cars which subscribe to that broker are considered consumers. Cars
can produce other data for use by other smart city components. For instance, cars use device-
to-device (D2D) communications to alert nearby vehicles and pedestrians of their presence and
potential traffic hazards. In a city scale deployment of smart services, data is generated at high rates,
which presents new challenges for smart city designers and developers.

Unfortunately, most of the generated data is wasted without extracting potentially useful
information and knowledge because of the lack of established mechanisms and standards that
benefit from the availability of such data. The main culprit is the lack of a large amount of labeled

2 Unlocking the potential of trust-based AI for city science and smarter cities - October 2019