AI and Security: Connected Twins

By Romi Mahajan, CEO ExoFusion and Paras Shah, CEO LAMR Group

Artificial Intelligence and Security are two areas of enormous importance, growth, and interest. While each has had tomes written about it, the convergence of the two spaces has received little attention. As we think about AI in particular, we are at the point where “rhetoric” is no longer interesting; we need to apply AI to particular verticals and disciplines in order to benefit from the clear advances being made.

For decades, Security was considered to be important, yet still unworthy of both a seat at the table and integration into what was thought of as “core business” functions. More recently, as organizations have digitized and cloud computing has proven to be the core intermediary between customers, partners, employees, and the various processes and experiences brought to bear by these organizations, security has been given the accord its due. Almost.

Many organizations still dither about the degree to which they need to make continuous investment in security. With the advent of AI, and its operationalization, this will change- not only because AI is the ultimate expression of digital transformation (which implies security) but also because the attack vectors multiply and acquire power via AI. We are in the era of “Wire Speeds.”

The convergence of AI and Security, then, can be seen on two fronts- The Security of Enablement and the Security of Disablement, the first confers enterprise advantage while the second offers a bounty of riches to the attackers.

Let’s pick three interesting areas from each-

Security of Enablement-the enterprise advantage

1. Access Control- before AI, UBA (User Behavior Analytics) was pie-in-the-sky. Now it is a reality. Enterprises can finally implement smart access control systems that dynamically grant and remove access based on user behavior.

2. Breach Awareness-AI has converted crude pattern matching into intelligent analysis of previously unimaginable volumes of data.

3. Internal Malfeasance- LLMs when combined with no code/low code tools enable security analysts to build threat surfacing models far more efficiently.

Security of Disablement-the attacker’s bounty

1. Ransomware- Attackers are increasing their ability to entice targets to divulge information with LLMs.

2. Spear Phishing- Attackers can use AI for precision in targeting- crafting messages that conform to culture, gender, regional, and role-based characteristics, even mimicking tone and style.

3. Complexity of attack modes- AI provides attackers with all the same data analysis improvements that benefit enterprises. Thus, the efficacy of attack modes has increased while the vectors have multiplied.

Indeed, the worlds of AI and Cyber-Security are intermingled and connected. Both areas must be understood to be essential parts of the “operating systems” of business. As IT and Business converge into a singularity, Security and AI are watchwords.

Where Security is a bit different, however, from other areas is that AI enables BOTH the enterprises looking to operate securely AND the bad actors looking to disrupt or create a payday.

AIsecurity
Comments (0)
Add Comment