Artificial intelligence (AI) is playing an increasingly important role in identifying software vulnerabilities, analyzing cyber threats and accelerating security responses, experts said, as the discovery of the high-severity PHP2Shell vulnerability highlights the need for faster and more comprehensive cybersecurity measures.
The PHP2Shell vulnerability, which received a severity score of 9.2 out of 10 under the Common Vulnerability Scoring System (CVSS), allowed attackers to execute code remotely and compromise web servers, exposing websites and platforms to potential security breaches.
Kubra Kalkan Cakmakci, an associate professor of computer engineering at Istanbul-based Ozyegin University, told Anadolu that AI significantly accelerates software debugging and can assess millions of potential errors simultaneously.
She said AI was able to examine multiple layers of the PHP2Shell vulnerability at once, a process that would take considerably longer for human analysts, who typically investigate only a limited number of steps at a time.
“AI was able to analyze multiple layers quickly at the same time when detecting the PHP2Shell vulnerability, while humans can only work on and think through few steps at a time,” she said.
Cakmakci stressed that effective cybersecurity requires a comprehensive approach covering every layer of a system, with organizations prepared to respond to critical security breaches.
“Ensuring cybersecurity against vulnerabilities requires a holistic approach addressing all layers, and organizations must have solutions in place for critical security breaches,” she added.
Beyond the challenges of identifying vulnerabilities, the potential exposure of sensitive user information remains a major concern.
Fatih Sensoy, a senior security researcher at cybersecurity firm Kaspersky, told Anadolu that the PHP vulnerability had forced many websites and platforms into maintenance and could expose plain-text session data and other sensitive information, including credit card details.
He said storing personal information in encrypted formats in separate databases could provide additional layers of protection, but warned that attackers who gain control of a frequently used website could potentially compromise everything accessible through it.
“Storing personal data in an encrypted format on a separate database on the server could add a few extra security mechanisms, but once a frequently used website falls into the hands of the attacker, everything from that point on is at their mercy,” he said.
Meanwhile, Sensoy cautioned against assuming that AI can independently identify every software vulnerability, explaining that initiatives such as Patch the Planet rely on human experts to direct the technology toward potential weaknesses.
He also highlighted the narrow window available to developers to address newly disclosed vulnerabilities, saying that preparing and releasing a security patch could take several days, while attackers might already be examining the vulnerability for ways to exploit it.
“An update is released to patch the vulnerability, but developing a patch could take up to two to three days, but in cybersecurity, even an hour is crucial — anonymous attackers are also on the lookout for these updates with the sole objective of figuring out how to exploit a disclosed vulnerability,” he said.
Sensoy said organizations therefore need rapid-response mechanisms to limit the damage caused by newly discovered security flaws, noting that such vulnerabilities have emerged in the past and will continue to pose challenges in the future.
“Such vulnerabilities emerged before, and they are being disclosed now, and continue to be disclosed in the future — when we face such incidents, we should have rapid response solutions in place to minimize the damage,” he added.