Can AI and ML technologies become the ‘Superhero-as-a-Service’ to help organizations manage the humongous troves of business and biometric data?
With the proliferation of IoT devices, organizations are collecting vast amounts of data from multiple touchpoints across their customers’ journey each day, to better provide personalized and improved customer experiences.
While the key challenge most organizations face is in managing the data and turning it into actionable business insights, what about the promise of behavioral data gathered from IoT devices to help organizations bolster their cybersecurity posture, using behavioral biometrics and authentication software to analyze and identify non-conformal use or suspicious activities in a user’s account, which can then alert IT and security teams to restrict access or prompt further investigation?
Can artificial intelligence (AI) and machine learning (ML) serve as a ‘Superhero-as-a-Service’ to help organizations more effectively and securely manage their huge silage piles of business and behavioral biometrics data?
CybersecAsia discussed the possibilities of leveraging the ‘superpower’ abilities of AI and ML with David Chan, Managing Director, Adnovum Singapore:
In what ways can AI and ML technologies add ‘superpower’ abilities to enterprises?
Chan: We like to think of AI and ML technologies as a ‘superhero’ for organizations, with superpowers at their service to help them manage their troves of data in a more effective and secure manner.
Different industries will identify different use cases for AI and ML, but in general, these technologies provide organizations with extrasensory ability, multiplying our human senses with the help of IoT sensors and correlate to outcomes, enabling organizations to conduct predictive maintenance based on the data gathered.
Well-oiled AI systems give the power of supersonic speed, allowing organizations to automate tasks and perform operations much faster than any human. This can come in the form of assistance at drive-throughs, extracting and classifying information from unstructured documents.
Referencing telepathy and precognition abilities, AI can analyze data in real-time, recommending the next best action, and predict potential scenarios. An example would be the transportation industry leveraging AI to predict train or bus capacity based on historical data to manage overcapacity.
Finally, AI grants organizations with x-ray vision – being able to detect and identify problems from volumetry, infrared images, heat and other indicators. This enables organizations to have a 24/7 access and view into the network infrastructure, continuously detecting and identifying any problems and threats to their network.
With data not being fully utilized for actionable business insights, how can organizations leverage AI and ML technologies to manage and analyze data more effectively? Please share some industry use cases.
Chan: Knowing what type of data and where it sits, how sensitive the data is, as well as identify the potential threat actors, are critical considerations for a well-organized data management process. This knowledge allows organizations to identify risk levels, prioritize their efforts, and prepare and execute effective data protection strategies.
AI-based discovery solutions are incredibly helpful in pinpointing where files are at any given point in time. Data classification solutions can then come in to inform how these data should be treated and protected, the policies required to be placed around it, and guide the prioritization of risk mitigation activities. AI and ML also has the ability to collect and assimilate huge volumes of data with speed, accuracy and scale then providing recommendations for how the data can be used to further business objectives.
Recently, Adnovum worked with Radio Television Suisse (RTS), a Swiss public broadcasting organization that was facing the challenge of managing its huge repository of images, audio and video assets. Without having visibility into where the data sits, they struggled to find the right data, and they required an army of documentation experts to help harness new insights and extract new information from this sea of digital content. A large amount of time was spent classifying the data manually and doing reports, and so we collaborated with them to develop an AI platform with audio analysis capabilities.
The AI platform provided them with better knowledge of the data they possess and enabled new digitalization use cases – helping them classify old records that were being digitized, collect gender statistics for radio programs, and improve other processing tasks.
How have IoT technology and biometric data been leveraged to bolster some organizations’ cybersecurity posture?
Chan: IoT is a familiar concept that has been around for a while, which allows businesses to gather data from the billions of connected devices and turns it into intelligence. But at Adnovum, we look at the extension of it – what we call Internet of Behavior (IoB).
With the help of AI/ML and data analytics, IoB synthesizes the data from users’ online activity from a behavioral perspective. This behavioral data can help security teams detect unauthorized access and suspicious activities by hackers, prompting them to activate security protocols at the earliest point of entry.
IoB and behavioral analytics capture more dynamic aspects of a user’s digital identity, such as their typing rhythm, mouse movement, geolocation, type of device used, the usual pages and links accessed, and even walking speed. Any unusual activities that stray from the norm will then trigger IT and security teams to restrict access to the account or prompt the real user to investigate.
Compared to adaptive security methods that constantly interrupt the user’s activity to verify their identity, behavioral analytics solutions work unobtrusively to continuously authenticate the user, which in turn facilitates a more seamless user experience. By capturing unique aspects of a user’s digital identity without requiring sensitive personally identifiable information (PII), organizations will also be able to better comply with data protection and privacy regulations.
Finally, user behavior is hard to replicate, and malware bots or fraudsters with stolen credentials stick out like a sore thumb when viewed through the lens of behavioral biometrics.