Upcoming WebinarsMore Webinars
Oge Marques explaining recent developments in AI for Radiology
Author of the forthcoming book, AI for Radiology
AI Business is part of the Informa Tech Division of Informa PLC
This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.
by Kumar Patel, Omnidya 10 October 2019
The global cost of cybercrime was estimated to be $600 billion in 2017, a significant jump from $445 billion in 2014. As hackers become increasingly organized and attack even the most high-profile companies, one thing is clear: the importance of cybersecurity cannot be underestimated.
Due to this, we have seen companies pioneering new protective setups, mainly by incorporating some form of artificial intelligence (AI). With AIís potential to adapt to new threats, eliminate errors, advance biometric logins, and identify threats at extreme speed - thereís no wonder why.
Insurance companies face increased vulnerabilities due to the nature of the data they handle - personally identifiable information (PII). PII can consist of a personís full name, personal identification numbers (passport, driverís license), bank account number, email, and other details. This data is an extremely valuable currency to those seeking to get hold of it. With such sensitive information circulating in customer portals, transactions, big data warehousing, and cloud data storage, insurance companies must make sure that each step of this cycle is protected.
The smart players know that this can only be done by designing and implementing a robust security architecture. But with a plethora of solutions available, what are the aspects to focus on in order to maximize the potential of an AI-powered solution?
As new innovations are constantly being developed alongside powerful systems, itís no surprise that companies can maximize AI by combining it with another technology. Blockchain is one such valuable ally. When working synergistically with AI, it has been proven to significantly improve data security. For example, a current blockchain network for AI data sharing ensures that the data is transparent, secure, and easily accessible.
The capabilities this partnership brings might be impressive, but itís hardly a thing of the future. From insurance giants like Allianz and Swiss Re to newer blockchain startups, security-conscious companies of all sizes are working to leverage solutions that include AI-powered fraud filtering, blockchain claim automation, fast document analysis, the advancement of smart contracts, and the simplification of information flow and payments between insurers and reinsurers. Even NASA is adopting blockchain to battle security, authentication, and privacy concerns.
While the 20th centuryís most valuable commodity was oil, the 21st century is all about data. By ensuring smooth data collection and management, companies can utilize AI to achieve things that were previously unimaginable. This isnít easy, though. Models that learn and adapt to the ever-changing threat environment and build predictive models require solid, accurate data sets. Immaculate data collection processes are a prerequisite for functional machine-learning setups. Automations can only analyze millions of data samples without a glitch, create correct patterns, and make optimal decisions if the data is impeccable.
Nowadays, companies can spend up to 80% of project time on data preparation, implementing a series of techniques to ensure execution doesnít fail. These techniques include data identification, profiling, sourcing, integration, cleansing, and preparation. While these processes might seem excessively laborious, they are essential. After all, identifying issues such as missing data in advance helps companies to find appropriate solutions in a timely manner without needing to reprogram an entire AI model. Interestingly, as AI advances, we are likely to see these processes become automated. Yes, AI will eventually request data itself without the need for human intervention.
For now, an effective way to prepare internal processes is to engage in mindful data collection. This term denotes the practice of considering how the data will be used before you even create or collect it. The power of the concept dwells in utility: how to optimize collection to serve your future goals in the most cogent manner? Mindful data collection puts the focus on data collection points within an organization and ensures that thereís no wasted time or data collected in vain.
Customer service can represent a big security hole as it forms a platform for the exchange of sensitive information including email addresses, passwords, and bank account numbers. Chatbots can help, however, because if built properly, they are technically more trustworthy than humans and less likely to commit errors.
For example, when resetting forgotten passwords or granting additional access, the presence of a chatbot decreases the necessity for third-party intervention. Eliminating traditional communication channels like emails can be useful, too. Since the beginning of the Internet, hackers have been utilizing phishing attacks, disguised as customer service representatives, to coax data from unsuspecting users. Added to that, emails tend to be extremely vulnerable; 91% of cyberattacks are thought to be launched with a phishing email, showing the necessity to either protect our inboxes or swap them with a safer alternative.
As long as proper security measures are in place, such as end-to-end data encryption, chatbots can be a major cybersecurity asset. They can be built to operate with authentication timeouts, making sure that any entry has a time expiration and canít be misused in the future. Likewise, they can incorporate two-factor identity verification or make sure that any sensitive PII is deleted immediately after the exchange takes place. Chatbots can also be set up to follow data privacy and protection policies, enabling companies to increase data security without burdening the customer.
There are few insurance organizations out there that havenít had to face a cyberattack yet, which means all the more reason to be meticulous when choosing a solution to keep you secure. Opting for a secure AI-based framework to combat threats is no longer a fancy add-on that only the biggest players can afford: itís a necessary investment that insurance companies of all sizes should consider.
Kumar Patel is the founder and CEO of Omnidya, the company that leverages artificial intelligence to create advanced customer-facing chatbots.
Author of the forthcoming book, AI for Radiology