Three Rules When Using AI to Add Value to Your IoT Smart Cities
AI adds tremendous value to IoT, but it is still a long way from having a technical, legal, economic and political environment to dramatically impact smart city IoT use cases. Government CIOs developing smart city strategies should exploit AI progressively to balance benefits and risks.
"Identifying use cases for AI" is ranked among the top challenges by Gartner's Research Circle members, implying a long way for AI to dramatically impact the smart city IoT use cases.
Cities failed to uncover the full potential of IoT and AI to maximize the business value as a result of siloed infrastructure and data.
The immaturity of AI technologies being applied to the fast-evolving IoT use cases multiplies the challenge of building internal confidence and winning support from city and function unit leaders.
To strategize and plan your introduction of AI into IoT use cases in smart cities, government CIOs developing smart city strategies should:
Manage and exploit AI in smart city implementations over time by prioritizing on high-value use cases for both governments and citizens in the next 24 to 36 months.
Push cross-organizational IoT facility and data sharing to create a synergy to both explore the potential of AI and deliver maximum results.
Boost confidence with a positive first experience of AI by starting small with low expectations and a one-off investment, and then creating a roadmap for future cases to stem from the initial one.
Strategic Planning Assumption
By 2020, in 30% of smart city implementations, artificial intelligence (AI) will become a critical feature, up from less than 5% today.
AI, as a type of analytics, offers tremendous capabilities to improve operational efficiency, drive sustainable competitive advantage and deliver business value faster (see "Artificial Intelligence Primer for 2018"). In smart cities, Internet of Things (IoT) that is combined with the capabilities of AI will enable applications to support much more complex data analytics and decision making. Cities can therefore realize new ways to run more efficiently and effectively, with optimized citizen satisfaction and experience. AI potentially makes IoT more useful to cities, governments and citizens.
The long-term vision of smart cities is full interconnectivity of city infrastructure, business processes and people. The incredible wealth of data generated through this full interconnectivity can be used to promote safety, efficiency, sustainability and economic growth. Smart city projects failing to make sense out of this huge volume of data emerges as low value and less sustainable. For successful smart city projects, government CIOs must establish new ways to leverage and monetize IoT data by embedding analytics throughout the IoT architecture, on the devices, at the edge, into the cloud, and across communication networks and IoT security.
AI and IoT are inextricably intertwined in three ways. The value and promises of both AI and IoT are being realized because of the other:
IoT data as the input to the AI system, particularly for training purpose
AI as the application of an IoT solution (for example, machine vision in video surveillance)
IoT and AI as a two-way system, with continuous interaction to each other's benefit, such as autonomous vehicles (see "A Framework for Applying AI in the Enterprise")
Gartner believes that 2018 will mark the beginning of a democratization of AI. This expectation offers government CIOs new opportunities to both experiment with AI in IoT smart cities and start to lay a foundation for successful piloting and exploiting it.
Treating and deploying IoT smart city data as an asset is still in its infancy. By doing so, cities are able to build up a competitive differentiator and a lighthouse effect across the industry. However, it's important to understand that AI won't be the "silver bullet" for all the city challenges or societal problems. And AI is not the only tool to exploit IoT data. In occasions in which real-time data needs to be handled and real-time actions need to be generated, techniques such as stream analytics can be easily and quickly deployed with no great skill required. One of the key decisions to make is to determine whether AI advances can increase business value more than traditional analytics.
In the meantime, the broad and rapid evolving nature of AI and IoT requires you, the CIO, to plan carefully. Using AI to add value to IoT smart cities is complicated and demands that government CIOs and stakeholders understand how AI can help and where the best place to deploy it is. As the government CIO, you shall walk a fine line between embracing and overplaying AI technologies' role in delivering business value for smart cities. There are three rules (see Figure 1) and six action items to consider to help realize the AI's benefit of delivering IoT business value faster, building up future-oriented IoT smart city use cases, while avoiding costly errors.
Figure 1. Three Rules of Exploiting AI in IoT Smart Cities
Rule 1: Manage and Exploit Over Time
AI still is a long way from having a technical, legal, economic and political environment to dramatically impact the smart city use cases. Both AI and IoT are still at their infancy of development cycles, featured with immature technology and still-emerging use cases. As indicated by the Gartner 2018 CIO Survey (see Figure 2), IoT and AI are definitely on most organizations' agenda; however, the majority of the organizations are at the early stage of exploring these two technologies.
Therefore, manage and exploit over time by introducing AI into smart cities incrementally, with the process for selecting ideal use cases and the starting point as thoroughly and well-substantiated as possible.
Figure 2. Both IoT and AI Are in Early Adoption
Prioritize on High-Value IoT Use Cases for Both Governments and Citizens
A survey with 83 Gartner Research Circle members indicates that, among 35% of the respondents, "identifying use cases for AI" was the top three challenges in exploring and adopting AI. It's impossible to recommend a single use case that is applicable for every city, because different cities have different priorities for their smart city projects. Among all the IoT use cases in smart cities, which keep evolving and expanding, ensure you give priority to those use cases of higher value. How can the value of use cases be defined in a smart city context then? There are some general principles to follow based on two key parameters: value that the project would bring to the citizens and value that the project would deliver for the governments. The real value of a smart city lies in the positive changes it makes to governments' operations and people's lives. Target a well-governed, well-regulated, principled AI-empowered IoT use-case roadmap that treats the public's interest and governments' returns as paramount.
Figure 3 demonstrates five key areas with examples of use cases that generate high value for both governments and citizens. It does not intend to give a definite answer for specific use cases for each city but, rather, provides a thinking process and methodology of screening out targeted use case(s). Of course, during this process, the city's key initiatives and objectives for a smart city shall be considered. Similarly, other factors, such as "ease of deployment," for a quick win shall also be considered, in addition to the parameters mentioned above.
Remember that IoT use cases keep evolving and expanding in smart cities. There are constantly new ones emerging that can be accessed using the same approach.
Figure 3. Prioritizing Areas and Use Cases for AI and IoT Deployment
The following real-life examples can be used to generate ideas for adding value to IoT use cases through AI in smart cities:
Impaired citizen-assisted living: Barcelona in Spain is pilot testing camera glasses with machine learning. These glasses analyze the image, facilitate safe navigation and crossroads, and guide people with visual disabilities, with the help of sound and the support of an application that is fully accessible via a smartphone that uses beacons.
Adaptive traffic light and automated accident response: The Traffic Police Bureau of Hangzhou Public Security Bureau performs real-time traffic prediction. The bureau uses video and image recognition technologies, to ease traffic congestion through real-time notification of an emergency needs handling and automated traffic system. It also provides citizens with real-time traffic recommendations and travel routes.
Crowd management: Saudi Arabia's Makkah Region Development Authority (MRDA) created a crowd-control system to increase safety and security of Hajj pilgrims. Data is collected via a wristband embedding identity information, special healthcare requirements and a GPS. In addition, surveillance cameras are installed to collect and analyze real-time video along the Al Mashaaer Al Mugaddassah Metro Southern Line (MMMSL), as well as in the holy sites, such as Great Mosque of Mecca, Mount Arafat, Jamarat and Mina.
Disaster prediction: The city of Seattle in Washington created RainWatch, which combines radar data with a network of rainfall gauges to monitor rainfall with a high degree of resolution, giving individual residents hyperlocaized warnings about a possible natural disaster.
Pollution control: The city of Beijing enabled high-resolution predictions of air pollution, based on weather (wind, humidity and temperature), topology, traffic and industrial activity.
Establish the Starting Point Based on AI Technology Maturity
AI is a collection of technologies that can be used in applications, systems and solutions to add specific functional capabilities (see "A Framework for Applying AI in the Enterprise"). Of all the breakthroughs in AI technologies, deep learning, natural-language processing and computer vision are leading areas of rapid technology advancement. These are the areas in which you need to lead your team to build knowledge, expertise and skills.
As a result of this technology advancement, IoT feeds AI-based applications to learn from the data, uncover new patterns or relationships, and recommend and take action. This virtuous cycle between IoT and AI at first is having the greatest impact on applications that exploit image and audio recognition.
Gartner expects the IoT-AI platform to go through three phases of adoption (see "AI on the Edge: Fusing Artificial Intelligence and IoT Will Catalyze New Digital Value Creation"). The same pattern applies to AI deployment in IoT smart city context:
Phase 1: Image- and speech-related IoT applications in the smart city
Phase 2: Smart city IoT applications that gain "transfer learning" from image- and speech-related applications
Phase 3: Address a wider spectrum of business needs in smart cities through advanced forms of machine learning
AI applications that have advanced most quickly benefit from visual and audio data, both of which are core outputs for many IoT endpoints in cities, such as surveillance cameras, wearables and drones. Video analytics will grow swiftly as an aspect of AI applications in government, as performance improves and as the applications that seek to employ such analytics results multiply. The majority of the real-life examples mentioned above were built through a strong leverage of video analytics. We expect there will be more deployments in areas of self-driving cars in vehicle-to-infrastructure (V2I) context, improved security solutions at high-risk venues (for example, airports and ports), supported living and healthcare.
Rule 2: Look Beyond Technology
The combination of IoT and AI isn't just about completing the loop of sensing, transiting, storing, analyzing and taking action from a technical point of view. It's really about a fluid experience and business value creation through optimized internal and external government processes. A technology-driven approach is not a prerequisite for smart city success. Cities should avoid a bias toward technological solutions by focusing on the "smart" rather than on the "city." It ultimately doesn't matter what the technology is called. What matters is that it can improve lives and deliver value to governments and citizens.
Ally AI Deployment to Drive IoT Productivity in Solving Government/City Challenges
AI plays a vital role in providing faster and more accurate IoT data analytics. This capability alone does not justify the investment without demonstrating enough real business value. This capability must be allied with expected business outcomes to drive IoT's productivity in solving government and city challenges. If there's something your city is already trying to solve through IoT, see if an AI approach would make sense or can optimize current solutions.
Only when AI is given the appropriate context and is designed to solve specific problems or challenges can it then help to harvest the full benefit of IoT data and fuel meaningful innovation. Figure 4 shows some contexts in which AI can be put into play.
Figure 4. Direct and Indirect Monetization Context of Smart City IoT Data
Manage AI Implementations Across Government Organizational Boundaries for Maximum Results
A city is a complex ecosystem of many verticals with industry-specific information and communication technology (ICT) systems. The complex challenge of bringing these different and overlapping systems together to solve urban issues gets more prominent and urgent to tackle in the face of IoT data monetization via AI. The organizational side of this AI-driven transformation is less obvious than the technological one, but it is just as important to anticipate. If you fail to adapt, the project will either risk being slow in generating results or fail to generate positive results. Look to organizational elements as much as technologies. Encourage working across organizational boundaries to maximize the investment in technology and to create a synergy to deliver better results. The creation of synergy will be achieved through access to a broad range of data from all aspects of a city.
The lack of coordination between the different domains of the cities — not only from a technological point of view but also in terms of how they operate — will result in duplication and higher costs. By looking at the possible use cases of AI in IoT smart cities:
The same networks of sensors and cameras used to improve transport efficiency can be used to save lives and fight against crime.
The same intelligent traffic lights used to improve traffic flow can be utilized by ambulances and fire trucks to get to the emergency scene quicker.
The same license plate recognition technology used to track parking can be used by law enforcement to find stolen cars and track criminals.
The sharing of facilities/infrastructures cannot be achieved without collaboration across organizational boundaries among government organizations.
In addition to the sharing of facilities, breaking smart city data out of its silos must be on the top of the agenda. Ironically, for the past few years, IoT has been quite successful building up data silos in smart cities. AI will play a vital role in terms of connecting and analyzing the siloed data. The value of your IoT data can be increased by combining it with other city data horizontally, to deliver additional insight and value. You are positioned to take greater advantage of what AI can offer.
Lead the initiative to oversee creating governance and interoperability standards, helping with procurement and contracts, as well as providing training on new applications and changing processes. As the implementation and data asset become more complicated in the long run, you should create cross-department teams to take over. It will be key that people who have a "seat at the table" are a mix of technical and business experts to ensure that smart city challenges are identified and addressed with the proper use of technologies.
Be reminded that changing the work culture and breaking the data silos may take time, given that these touch and impact the interest of different government departments. Therefore, an IoT data strategy should be in place by treating IoT data as a strategic asset, which can be decoupled from applications so as to be shared and analyzed across programs and departments. In the meantime, make sure you build a linkage between data sharing and departmental performance, so that different departments are motivated to share their data.
Rule 3: Build a Positive First Experience
For political reasons, governments are more failure- and risk-sensitive than most of the other verticals. Many decisions heavily rely on a balance between inherently short-time politic horizons with long-term city strategies. With such an evolving topic as AI and its role within IoT smart cities, it's critical that the first experience for the city is positive. Rushing into an implementation without realistic expectations and risk-control measures from the start only sets up a deployment for failure. Build internal confidence with a positive first experience of AI applications, and then communicate that positive experience within the business. This strategy will help you to create a virtuous cycle and drive long-term success by expanding into other business areas and encouraging collaboration on how AI can transform processes.
Set Realistic Expectations on AI With Municipal and Departmental Leaders
The market's hopes, fears and expectations for AI technologies have been far too high (see "Hype Cycle for Artificial Intelligence, 2017"). Managing city stakeholders' expectations about what AI can or cannot do is becoming increasingly important. With no cookie-cutter solution and roadmap, most AI projects in smart cities will likely be a matter of trial and error. Disagreements and disputes are unavoidable during execution. Overly optimistic expectations can compromise your reputation, while overly pessimistic ones can compromise the competitiveness and differentiation of the city.
Balance your goal of winning support against the risk of setting unrealistic expectations. Realistic expectations can be set only when all related parties regard AI judiciously and are realistic with what it can deliver to drive up IoT deployment value. Provide city leaders with enough knowledge that they are no longer susceptible to AI-related media and vendor hype. And educate different government functional unit leaders on the most relevant facts about AI to best-inform their decision making. Be careful to avoid giving them too much actionable content so they don't go around the IT organization or override the IT organization to force-fit the technologies for both appropriate and inappropriate IoT use cases.
Begin small with a low expectation, and create a roadmap for the future in which more business cases can stem from the initial one. Set up milestones, and ensure that all stakeholders agree on the long-term goal of incorporating AI in IoT smart cities. The goal is to achieve expected business outcomes through monetizing huge volume of in-flood and accumulated data collected by IoT devices (see "Prepare to Monetize Data From the Internet of Things").
Be open-minded to understand, evaluate and adapt what other cities and organizations are doing, both inside and outside the city context. How other organizations are evolving their approaches to exploit IoT data with AI, either positive or negative experiences, might shorten your learning curve and inspire ideas in unexpected ways. Gartner presents research (as an example, see "Lessons From Early AI Projects") on this topic.
Manage Risks of AI Technology Immaturity and IoT Data Sensitivity
Despite the significant benefits brought by AI, the initial costs and risks can be very high. For most cities, the deployment of AI is a first and a learning experiment. Do carefully balance costs, risks and benefits, to limit the scale of unwanted side effects.
The risks of AI come mainly from technology and market immaturity. Although many core AI technologies are proven, the market for using those technologies, in general, is still in its infancy. Be prepared for rapid product and solution change, which is too great and fast to lock into any one platform or vendor at the moment. Plan for quick time to business value to counteract rapid obsolescence. Look to leverage cost-effective compute resources via a communications service provider (CSP) wherever applicable. Consider using off-the-shelf AI algorithms and models, rather than investing in specialized in-house development. Consider starting projects in the public cloud for both a low startup cost and the ability to scale to accommodate future requirements, as well as for an open architecture with multiple machine-learning frameworks and toolkits.
Another potential risk comes from the sensitive nature of the information involved in the process of exploiting IoT data (see "Data Risks in the Internet of Things Demand Extensive Information Governance"). This issue is particularly prominent in a city context. Make sure to put legal and ethical issues on the front burner, refining existing data governance disciplines and capabilities to address the increased risks of higher-sensitivity IoT data by extending policies to cover the "things" data.
The analysis and advice provided in this document are built from constant scanning of the market, as well as the aggregation of analyst experience and ongoing interactions with end users and technology and service providers. We have used a range of sources, including the following:
The 2018 Gartner CIO Survey was conducted online from 20 April 2017 through 26 June 2017 among Gartner Executive Programs members and other CIOs. Qualified respondents were the most senior IT leader (CIO) for their overall organization or a part of their organization (for example, a business unit or region). The total sample was 3,160, with representation from all geographies and industry sectors (public and private). The survey was developed collaboratively by a team of Gartner analysts, and it was reviewed, tested and administered by Gartner's Research Data and Analytics team.
AI development strategies research was conducted via an online survey from 5 April through 21 April 21 2017 among 83 Gartner Research Circle members — a Gartner-managed panel composed of IT and business leaders. The survey was developed collaboratively by a team of Gartner analysts, and it was reviewed, tested and administered by Gartner's Primary Research team.
Interviews with cities as early adopters of AI at various stages of implementation from December 2017 through February 2018.
Some opinions expressed in this article may be those of a guest author and not necessarily Analytikus. Staff authors are listed in https://www.gartner.com/doc/reprints?id=1-4XYENKG&ct=180501&st=sb&mkt_tok=eyJpIjoiTWpZeFlUWmhOek01WmpZMiIsInQiOiJpRlFnaEJHd2k1dkVTUTdEaGU1YTBQYkZHMkdyQklcL043dU92YUlFRm95RlFlTWZ4RThOdTA4OTU5bHZUY1VtZitcLzRyTTA5aXRyelJQazV1ZzdxWWlybjdvWUNNdUp3UWJYVTh4TE9FUlBjTkpcL3lYOW1QSDdmM0FKQjdlckdaV1lIZTVDVzdFcmIwb2xwV2JlY211aEE9PSJ9