As with most new technologies in security, there are quite a few challenges that security agencies must overcome to make the most of predictive analytics. As with most new technologies in security, there are quite a few challenges that security agencies must overcome to make the most of predictive analytics. According to Adlan Hussain, VP of Marketing at CNL Software, these challenges can be broadly categorized as technical and political.
Cities have invested in security technology for decades. They often have a wide range of disparate systems, some of which are new purchases that have been made for a specific problem and others are old systems that still provide value to a city.
“The challenge is to bring these systems into one platform that allows all assets to be seen from one view,” Hussain said. “The proprietary nature of these systems and the differences in the technology they are built on often make it difficult for them to talk to each other.”
Maurice Singleton, President of Vidsys added that one major challenge being faced is, there is still a lack of proper understanding in regards to how the massive amounts of data collected from different kinds of sensors can be put to use. To overcome this, there is an increased effort being made to raise awareness on how to leverage analytic tools.
Analysis and model programming
Predictive analytics use numerous smaller models to create one overarching model that detects anomalies. Errors in coding within these models or incorrect modeling can lead to incorrect analysis. These errors lead to incorrect reports, leading the security team to waste critical resources chasing potential threats from incorrect conclusions generated by predictive analytics.
“Periodically running the model against scenarios where an outcome is known and understood will help benchmark the model and ensure proper analytic functionality,” to Brian Schwab, Founder, and Principal Consultant at S3SDC.
Another issue is that algorithm-based models are static, but the real-world security environment is dynamic. Models for predictive analytics may be accurate at one point in time, but changes in threat behavior and other activities require models to be continuously updated using correct, relevant data.
“Security teams must have an expert on staff who is knowledgeable in how the model operates or contract with a third-party provider to update the models on a continuous basis to ensure that new operating parameters are identified, models are adjusted and proper output is maintained,” Schwab added.
Cities are made up of multiple agencies, government bodies, emergency responders, private entities and other stakeholders. This creates its own challenges as a holistic approach to security is needed for predictive analytics to be effective across a city. Sharing of assets such as video surveillance cameras and data such as blacklisting is vital for city-wide security.
“The last few years have seen more public/private partnerships, which have been enabled by MOUs between stakeholders,” Hussain said. “These agreements allow the effective sharing of data in the event of an incident. Technology can ensure that permissions are only given when needed and can help to ensure that data is not misused.”
Taking prompt action based on data collected is another aspect that authorities need to look into. Once armed with the data, city officials must act properly to protect people and assets. Failure to act once data indicating a potential problem has been received means the city has not met its fiduciary responsibility to guarantee security.
“The city has vicarious liability for any incident that occurs which could have been prevented,” Schwab said. “Ensuring that legal departments are
an integral part of the planning, deployment, and operation of these systems will lessen the legal burden a city will face after an incident has occurred. In addition, conducting adequate physical and cybersecurity risk assessment will highlight areas, where vulnerabilities exist and implementing mitigations for those vulnerabilities, will reduce the instances of liability that can be experienced after an incident.”
Implementing smart solutions requires considerable financial investment from governments and private players. Lisa Brown, Senior National Director of Municipal Infrastructure and Smart Cities at Johnson Controls and Donal Sullivan, VP, and GM at Johnson Controls Ireland said that funding and leadership were the main challenges.
“According to our Smart Cities Indicator Survey, 36 percent of respondents in North America identified unavailability of appropriate financing options as the top financial barrier. In the case of leadership, 18 percent of cities identified lack of city leadership as a barrier while another 18 percent identified a lack of state or federal government support,” they said. “In addition, in order for a municipality to implement a cyber and data governance effort, they need to work together, and this often comes with its own challenges, as the local government has historically been very siloed.”