Artificial Intelligence and Data Security

By CIOReview | Wednesday, January 17, 2018
234
341
67

While the larger talk surrounding AI only revolves around its positive and transformative impact to the world, there is hardly any talk about its implications. The existing mainstream critique of AI is more often tainted by the reductionist Hollywood imagination of the machines taking over or the end of human race, thus, neglecting the more immediate and larger repercussion of AI, namely, its implications to data security.

A classic case to substantiate the data security challenge is AI’s impact on the patient health records. As AI solutions are developed one-sidedly, aimed at improving patient engagements, or decision making or automating clinical processes, it has just come to open up more platforms and portals to breach the sensitive health data. The lack of thought to develop AI based security solutions has emerged as an instigator to more serious and damaging cyber attacks via ransomware, viruses, malware that are facilitated with more easy avenues to attack.  

Recommended: How workspace analytic allowing us to change the way we work?  

Another major danger revolving around AI and data security is the growing tendency to overwhelmingly shift the data security responsibility to AI. Debating purely on the lines of automated solutions for cybersecurity, and thus, failing to recognize the need for effective human involvement. A recent study by MIT’s AI lab, argues that a judicious combination of human experts and AI-based systems might be the ideal solution to detect and eliminate threats. It argues for humanly supervised machine learning security, which makes sense for there is always room for the system to go rogue or the regulating technology in-turn to be effected, posing more serious repercussions than the absence of an automated solution.