Cybersecurity leaders are sounding the alarm after a serious vulnerability was discovered this week. They predict that in the coming months, artificial intelligence (AI) tools will uncover even more fundamental weaknesses in the global internet infrastructure, reports UNN, citing The Guardian.”I no longer think we will be able to fix all the bugs,” Sam Altman said on Thursday evening, joining other industry leaders who have voiced similar concerns. Against the backdrop of events already dubbed “Black October” (during which researchers used AI to identify critical flaws in a programming language used by millions of websites and which may have remained vulnerable for years), technology sector representatives are warning that AI will be used both to find and to exploit vulnerabilities in security systems—and on an unprecedented scale and at incredible speed.AI agents were left alone without any tasks – the result stunned researchers23.10.26, 10:09 • 21752 viewsAccording to Jason Clinton, Anthropic’s deputy chief information security officer, a “tsunami of vulnerabilities” is coming. “There simply will not be enough time to patch critical infrastructure,” Clinton wrote on the social network X on Thursday evening. Marc Benioff, CEO of Salesforce, agreed, calling the situation a “red alert level of threat.”Matthew Prince, CEO of Cloudflare, said that AI is capable of finding and identifying security gaps in ways and at a speed that will have catastrophic consequences: cybersecurity professionals simply will not be able to respond to these threats in time. “It will feel as though the sky is falling to Earth. It will feel as though there is nothing we can do.The next two years will be extraordinarily, simply unbelievably, packed with the emergence of new tools that have limitless patience for searching for vulnerabilities across the entire internet,” he wrote in a post on the social network X.Although cybersecurity experts have been warning for several months about a “vulnpocalypse”—a situation in which artificial intelligence makes it significantly easier to identify security flaws—it was during “Black October” that researchers managed to use AI to discover such a fundamental and critical vulnerability, which was announced on October 21.Serious vulnerabilities were discovered in the PHP programming language, leaving millions of websites built on it potentially vulnerable to hackers for years. According to some experts’ estimates, 70% of all websites on the internet were at risk of being hacked.However, although AI will help identify vulnerabilities at unprecedented speed, technology industry leaders fear that malicious actors will use it just as quickly to exploit these flaws.Attempts to exploit the PHP vulnerability have already begun.After information about the vulnerabilities was disclosed, operators of government and private websites in the United States and the United Kingdom temporarily suspended their operation to address the flaws. The U.S. National Security Agency (NSA) issued a warning recommending that administrators of all websites running on PHP install the relevant updates immediately. The NSA noted that malicious actors could try to exploit these vulnerabilities to gain unauthorized access to servers.Zuckerberg wants to create a giant online network of AI agents22.10.26, 11:32 • 7116 views”Such an access vector could allow malicious actors to steal confidential data, disrupt services, or penetrate deeper into the network,” the warning said.In a post on the social network X, Prince said that the security company CloudFlare had recorded attempts by thousands of users to exploit PHP vulnerabilities as early as 10 hours before the official announcement.”Within 45 minutes of the announcement, the number of such attempts had risen to tens of thousands; after two hours, to hundreds of thousands; and now we are talking about millions,” he wrote.In their blog, researchers from Wiz, a platform focused on securing artificial intelligence systems, noted that this was “one of the most serious vulnerabilities ever discovered in PHP, and possibly one of the most dangerous in the history of software.””These fundamental flaws had existed for decades, but demonstrating their exploitability required an extraordinarily complex chain of actions—at the limit of what human researchers are even capable of carrying out in practice,” they wrote.
——————————————————-
Click Here For The Original Source.
