Conversational Interface for the Connected City of the Future

Angel, the virtual LA City assistant (Bot vector: Oksana Latysheva)

Imagine a city of the future, full of IoT, automated vehicles, and robots. Now imagine you are walking in the park and suddenly in front of you is an elderly man who got a cardiac arrest — an emergency situation. You, a good samaritan, approach him to give the first aid, your phone has run out of battery, and there are no other people around you except these IoT (a smart street lamp or a connected trash can). How would you ask it for help? Or at least ask it to contact the emergency service?

Perhaps this scenario seems not unlikely in the ever-increasing urbanized cities where more and more people have a smart phone. However, the need ability to communicate with “smart infrastructure” introduces a lot of exciting possibilities.

First, the built environment that we know will consist mostly of these objects. One report predicts that the number of IoT will be doubled in 2020. Therefore, it is important to think of way we can interact with them. Second, from the perspective of city government, service delivery is the key to the livelihood of its people. The urban environment, infrastructure, facilities, and services, depending on how they are planned and built, can impede or enable access, participation, and inclusion in society, especially for accessing the City’s services. Therefore, providing city services that are accessible to everyone, regardless of one’s technological literacy is important for achieving an inclusive city. The conversational interface offers this convenience because it does not require users to be good at operating the device, in fact, you can even create a device that receives conversational input on your own.

HAL 9000 takes conversational input, no additional knowledge required

Additionally, there is also concern about data integration when we talk about IoT and the connected future. Our government structure is divided into many jurisdictions introducing the potential for miscommunication and replication of services. Many cities around the world trying to tackle this issue by leveraging mobile apps and web apps to allow people make service requests in a centralized fashion. While this strategy is smart it is limited; as this trend is growing, people’s willingness for downloading new app is shrinking, and therefore make this strategy ineffective for improving citizen engagement.

Screenshot of the App from iTunes Page (source: iTunes)

The City of Los Angeles service portal, the MyLA311, was built to address the same concern, that the City should confront the need of a centralized system for citizen engagement as well as to prepare itself for the upcoming wave of technology. Sadly, even though the system itself has a great range of functionalities, the rate of adoption of the application is quite low. Since its release in 2013, only 20% of 311 requests are made through the mobile application. For the system to live up to its promise, the City should find a better strategy to improve the system adoption. For all the reasons explained before, the conversational interface might be the key solution to this problem.

This writing recommends that the City of Los Angeles explore the feasibility of employing conversational interface for its service portal system. In addition to increasing adoption of 311 which will help the City become more responsive to citizen needs, the conversational interface will also help the City better prepared for the generation of technology to come, where infrastructure will be all connected and and the form of civic engagement might take place in both physical and cyberspace.

Implementation Strategy

Here are the implementation strategies that might help the City moves on integrating the conversational interface into the city service request and delivery system. This recommendation is based on the writer’s knowledge of the currently available conversational technology, in this case, chatbot and virtual assistant (e.g. Google Assistant, Alexa), and without getting into the technical side of the implementation.

First, the City of Los Angeles’ innovations over the MyLA311 portal should act as a cornerstone for developing this new system. It is important to note that the MyLA311 has provided a precedent on how a centralized city service
system might work. Therefore, this new system should be understood as a strategy for extending the reach of the public services rather than a
completely new system. In short, the conversational system will not change how the portal works, but will replace the ineffective smartphone application.

Second, the current MyLA311 is managed by The Los Angeles Mayor’s Office and the Information Technology Agency (ITA) with most of the work focused around managing and depositing the citizen requests into a central repository. This system might be sufficient when dealing with the current volume of requests. However, when the number of requests surpassed the Office’s capacity, either in volume or on jurisdiction, it is preferable to have a dedicated line that handles such requests.

Finally, partnership with private technology companies can help the City maximizes the use of available personal devices for improving the emergency responsiveness. Having access to personal sensors, the intelligent system can sense the apparent emergency situation based on case scenarios. Additionally, accessing sensors of a personal phone through the API can alleviate the user’s skepticism regarding government surveillance. The API acts as a sandbox for developers (the government), where users can still retain control over the shared data.

Timeline

Today (0–3 years)

• Impact analysis to study the budget, legal framework, risk analysis, and the privacy principles.
• Cooperation building, which include inter-agency meetings within the city level to identify types of city services to be included in the chatbot system and also to set up the most optimal pipelines for forwarding the service requests to the back-end users (city departments).
• Infrastructure building through public and private partnership, which include equipment and system vouching to identify the best form of chatbot technology available to serve the purpose.
• Professional development for teams or agency that will work to monitor the traffic of request and provide technical support for the back-end users (city departments).
• Capacity building for the workforce at the back-end of the system (city departments), to familiarize them with the advantages and limitation of the system.
• Beta-testing the chatbot to assess the potential fallacy of the early stage of the system, which based on the sandbox deployment to minimize externalities (e.g. focusing on one type of chatting app, certain geographies, or demographics). One of the potential environment for the beta-testing is within the government itself by allowing the chatbot to assist tasks done by the city officials. This also can serve as a training environment for the machine learning algorithm.

Tomorrow (4–6 years)

• Building a library of protocols for specific issues resolved during the first years of deployment, which including the evaluation and improvement of the previously adopted privacy principle.
• Establishing feedback loop for services and infrastructures, which means continuous evaluation and improvement on the use of the chatbot system from the back-end users (city departments).
• Incremental expansion of the services to the other kind of platform (e.g. every chatting application, bus stops, parks) and its integration with the other kind of personal assistants (e.g. Amazon Echo, Google Home, automated vehicle).
• Finalizing the legal and institutionalized framework needed for the whole city and demographic deployment of the system.

Future (7+ years)

• Whole city deployment of the system.
• Connected, responsive, and user-friendly infrastructure where the chatbot become a personal assistant for every city-related services (e.g. transits, street infrastructure, emergency services, licensing, permits, and public safety).
• Leveraging the data to better allocate city’s budget for services outside of the chatbot scope, which can include the optimization of the City’s plan for economic and community development, transit management, housing projects, and many other city tasks.

In the end, even though the goal of the current system is only to focus on optimizing the city services and for establishing the City’s real-time feedback loops, in the long term the system should serve as a backbone for embracing the ultimate smart city, where city services will have to deal with fully connected urban infrastructure. Conversational interface will help humanize these smart infrastructures and help people interact directly with it.