Intelligent CIO LATAM Issue 10 - Page 83

Attackers can attempt to control the data sets that train the AI , for example by subtly altering parameters or modifying scenarios to avoid detection of underlying data exploits . Similarly , pattern recognition can be used to identify access points for injectables for remote execution at a later date , or even to improve social engineering by targeting workers at their most vulnerable moments . A simple mention on a social media website about network maintenance could alert cybercriminals to a potential weakness .
At the same time , AI can also be deployed for protection . The best line of defense is usually to retaliate in the same way . AI is already being adopted in data analysis and network monitoring , where it is used to determine a baseline of normal behavior and identify inconsistencies of different kinds , such as unusual traffic patterns or anomalous server access . As the algorithm learns and progresses , predictive analytics can be implemented to detect such intrusions early on , while implementing defensive responses and triggering supervisory alarms .
As technologies like AI rapidly evolve to integrate into the industrial passageway , cybersecurity issues will continue to be a key area of concern . Security professionals must assume that AI and other technologies can and will be used for criminal gain . Global cybercrime is expected to inflict US $ 6 trillion in total damage this year , increasing to US $ 10.5 trillion annually by 2025 . A significant percentage of those attacks are likely to affect industrial organizations . A comprehensive approach that anticipates and predicts cyberattacks can protect organizations from security problems .
Criminals use a wide variety of methods , from commonly used techniques such as phishing and password hacking to more sophisticated operations .
www . intelligentcio . com INTELLIGENTCIO LATAM 83