Skip to main content
CybersecurityHacking

Exploiting Existing Vulnerabilities Fuels Cyberattacks

Modern office interior with computer workstations and a blurred laptop screen in the foreground.

"The majority of attacks that are happening right now are not because attackers are finding more vulnerabilities. It's because they're exploiting more existing vulnerabilities," said Jonathan Zanger, CTO of Check Point.

Jonathan Zanger: AI speeds exploitation, but not discovery

Zanger told iTNews Asia that AI is changing the tempo and scale of attacks by helping threat actors identify vulnerabilities, develop malware and run phishing campaigns faster. AI has also "lowered the barrier to entry for cybercriminals, allowing them to carry out attacks with fewer resources," he said. Crucially, Zanger cautioned that the rising volume of attacks does not equate to the discovery of many new flaws; instead, attackers are increasingly weaponising long-standing, existing vulnerabilities.

Compromised credentials, phishing and the "long tail" of gaps

According to Zanger, many breaches still begin with familiar entry points: compromised credentials, outdated systems and phishing. He warned organisations that "people are overlooking the long tail of security gaps, which is usually how attackers get in." Leaked credentials that remain active, weak passwords and internet-connected devices lacking adequate protection are all cited as persistent problems. Zanger added that AI enables attackers to process information and move through compromised networks more quickly, amplifying the impact of these basic failures.

Practical controls: patch management, network segmentation and threat prevention

For organisations with legacy infrastructure or constrained security budgets, Zanger urged a pragmatic approach rather than wholesale replacement. He recommended improving patch management, tightening network segmentation and making better use of existing threat-prevention capabilities. He singled out the task of protecting devices that cannot be patched immediately — common in manufacturing facilities and hospitals — and suggested examining how attackers could reach those devices and using existing network infrastructure to block those paths.

Prioritise exposure, adopt zero-day protections and design secure AI

Zanger set out three priorities for security leaders. First, understand actual exposure and prioritise remediation based on how attackers could exploit vulnerabilities, rather than trying to fix every issue at once. Second, strengthen defences against previously unknown attacks — he pointed to web application firewalls and urged organisations to "look beyond solutions that rely solely on known attack signatures and consider zero-day detection capabilities." Third, build security into AI systems from the outset: "Design the system initially with security in mind, so you wouldn't have the discussion in a couple of years of how to make it secure."

He described three stages of AI security: identify where AI is being used; protect the channels through which data enters and leaves AI systems; and monitor agents while they operate. On autonomous AI agents that can access enterprise systems and act for employees, Zanger warned that conventional tools "were not designed to manage agents that interact with multiple systems and make decisions independently," and that organisations therefore need "an external security mechanism to monitor an agent's purpose, permissions and actions, and block activity that falls outside its authorised scope."

What this means for education, telecommunications, and manufacturing

  • Education: As one of the most targeted industries globally and across APAC, educational institutions must prioritise credential hygiene and rapid exposure assessments to limit the long tail of vulnerable systems.
  • Telecommunications: Because attacks on telecom providers can expose customer information used in follow-on operations, telecoms should emphasise zero-day detection, segmentation and stronger protection of data flows.
  • Manufacturing (hardware and semiconductor): With many unpatchable or legacy devices in operational technology, manufacturers need network-level controls and exposure mapping to prevent lateral movement to critical assets.

Zanger also highlighted continuous threat exposure management (CTEM) as essential. As organisations deploy new applications, onboard employees and introduce technologies, their attack surfaces change; CTEM, he said, helps organisations view networks from an attacker's perspective, identify weaknesses and prioritise remediation instead of being overwhelmed by the volume of alerts and vulnerabilities. He noted that the number of cyberattacks observed globally had increased by more than 70 percent over two years, underscoring the financial incentives driving cybercrime.

The practical takeaway from Zanger's remarks is straightforward: while AI raises the speed and reach of attackers, the most effective immediate action remains shoring up long-standing weaknesses — patching, segmenting and using existing prevention tools — while designing AI systems with built-in security and human oversight. Organisations must address "both sides of the threat" he said: strengthen existing controls and introduce safeguards as AI becomes more embedded in operations.

Original story: iTNews Asia