Vibe hacking and agentic SOCs: Google Cloud Security exec on evolving threat landscape

As cyber threats evolve in the age of generative and agentic AI, attackers are no longer relying solely on technical exploits. A new class of risks is emerging that takes advantage of how humans interact with AI systems. One such phenomenon is vibe hacking.

At the simplest level, generative AI has made coding dramatically easier. For instance, one could ask an AI tool to build a complete e-commerce business—including the payment gateway, and it could generate the entire stack. AI is lowering the barrier to building sophisticated systems. But on the flip side, if coding has become easier, so has hacking. 

“This is where the emerging concept of vibe hacking comes in. In a world where businesses are rapidly adopting AI tools, hackers have discovered new ways to exploit them. Instead of targeting only technical vulnerabilities, vibe hacking manipulates tone, trust, and emotional cues in AI-driven interactions,” said Jagdish Mahapatra, Managing Director, Google Cloud Security, JAPAC, in an interview with TechCircle. 

Mahapatra joined Google Cloud Security’s team this year, where he is leading the APJ region that comprises Japan, Australia, and Southeast Asia. A security veteran with stints in companies like Cisco, McAfee, and CrowdStrike, Mahapatra spoke extensively about the evolving landscape amid developments in AI technology.

“The danger is that people rely heavily on these AI systems. If manipulated, the consequences could be far more damaging than traditional attacks. The key takeaway is this: the threat landscape has shifted,” he said. 

Adversaries are moving faster than ever, stealing data and disappearing before traditional defenses can react. This makes rapid detection, investigation, and response critical. Mahapatra suggests that organisations must modernize their security operation centres (SOCs) with two essentials: cloud-native technology that operates at cloud speed, and AI-powered tools that enable seamless collaboration between humans and intelligent agents. “Threats don’t discriminate, both startups and legacy enterprises alike need this transformation,” he said.

The scope of AI agents has expanded to include its relevance in the security operations environment, Mahapatra noted. Called Agentic SOC, these agents can take on a significant share of the repetitive, high-volume tasks. “This relieves human analysts, who are increasingly stretched, and allows them to focus on higher-value, complex investigations. The future SOC will be built on this collaboration between humans and AI agents,” said Mahapatra.

Google recently demonstrated a live cybersecurity agent called Big Sleep, developed by Google’s research lab, Google DeepMind. Big Sleep is designed to detect vulnerabilities in software and was able to identify weaknesses on the fly that were previously undetectable by humans, said Mahapatra. 

“Think back to the Log4j vulnerability from a few years ago. If AI agents like Big Sleep had been available then, they could have flagged the issue early, preventing developers from releasing a version with such a critical flaw. That’s the kind of proactive defense agents make possible,” he adds.

The Log4j incident refers to a critical cybersecurity vulnerability discovered in Apache Log4j, a widely used open-source Java logging library, in December 2021. The flaw allowed attackers to execute arbitrary code on affected systems simply by sending a specially crafted log message.

India market

India has long been a strategic market for Google and continues to rank among the fastest-growing globally. It sees strong adoption across key sectors such as BFSI, government, and IT services. With its young population rapidly embracing digital products, India is uniquely positioned for cloud-native business models. This brings clear advantages, but also introduces new security risks.

“What makes India distinctive is the juxtaposition of high digital adoption with, at times, limited awareness at the broader population level. This places significant responsibility on organizations, whether banks, telcos, or IT service providers with global clients, to ensure robust cybersecurity. As a result, the demand for security solutions continues to rise across all verticals, and we’re seeing growing engagement with customers across the board,” said Mahapatra.

At one end of the market, cloud-native digital-first companies find it relatively easier to embed security into their systems from the start. Their agility allows them to move quickly and adapt modern practices seamlessly. On the other end, traditional enterprises—such as banks in India—have also been highly proactive in implementing strong security measures. Banks, in particular, have long partnered with the right providers and adopted robust frameworks to safeguard their operations, he noted while describing Google Cloud Security’s clientele in India. 

“Today, we are seeing significant cloud adoption across both banking and other traditional sectors, creating an opportunity to strengthen security further. Security itself is becoming inherently more cloud-native. A decade or so ago, security was designed primarily for on-premises environments. Then came the shift to cloud-native solutions built for on-prem, and now we’re in an era of cloud-native security that protects both cloud and on-prem workloads,” he added.