AI, machine learning, Technology / By Gal Ben-Tovim / July 14, 2019 / AI, cyber, cybersecurity, Technology

Hudi Zack, the Chief Executive Director of the technology unit in Israel National Cyber Directorate, said during his presentation at the Tel Aviv Cyber Week AI conference last month:

“great benefits come with great dangers”

Hudi Zack, the Chief Executive Director of the technology unit in Israel National Cyber Directorate

people should be aware of the risks brought by Artificial Intelligence (AI) while seeking opportunities from it.

 In addition to the multi-facet risks and cyber exposure that AI technology has generated, enterprises can be infiltrated by “turbo-hackers” (AI-driven hackers). This can cause havoc amongst critical AI systems and destroy the flow of work in the decision process. 

Cyber resiliency is an important AI system attribute and is a part of the complex web of trust, people should start to think about this issue and come up with a solution to it. Below are two inspirations that governments, enterprises, and individuals should integrate. 

Following the introduction of AI and cybersecurity, Hudi Zack then talked about privacy-preserving machine learning (PPML) and its application of homographic encryption (HE). This is one of the two techniques that should be used when untrusted parties are using machine learning together. This technique will be estimated to have a market value of 22% of the US GDP. For example, when using the homographic encryption, retailers could pull data without sharing it with banks. Third-party AI companies could then apply machine learning while the data remains encrypted. Banks could then collect personal data with quantifiable privacy protections. 

One of the use cases is AI companies as they could help banks to improve their business model with the data from retailers. These use cases promote the concept of open banking and the use of open APIs. Although there are difficulties deploying deep learning on homographic encryption, Hudi Zack made his point by saying:

“advancing both AI and privacy is not a zero-sum game and no single technology solves privacy”.

Hudi Zack, the Chief Executive Director of the technology unit in Israel National Cyber Directorate

Another speaker Dr. Dorit Dor, Vice President of products at “Check Point”, mentioned that AI and Cyber need to be in a close collaborative relationship. Cyber defense, based on Artificial Intelligence, can identify negative motives, by looking at the data stream sample and the attack.

For example, data scientists can leverage a Chief Information Security Officer (CISO) to have a more accurate analysis. AI provides a data label from links to understand network activities from previous behaviors and tag them to label the whole cluster. It also extends the analysis from different vectors by enriching URLs in different scenarios and adding users data to enterprise data. Therefore, a company can obtain a considerable diversity of data samples instead of similar ones to learn an effective methodology that really captures the core.

With new generations and a new threat concept of cyber attacks coming up in the future, “Check Point” can also identify malware DNA. By looking at the content from both data analysis and behavior analysis and classifying the distribution, “Check Point” thus is able to proceed to the threat analysis, come up with solutions and provide a safer environment for cyber. 

This is proof that AI is a tool that can be used for the good, so every sector should be aware of AI-driven attacks and leverage them for cyber defense measures in the meantime.

Leave a Comment

Your email address will not be published. Required fields are marked *