The AI era has made many things easier, such as coding or research. But it has also given hackers more tools for cyberattacks. Google has now warned that hackers are using large language models (LLMs) to identify software flaws and develop exploits without a human in the loop.In its report, titled “GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold,” the Google Threat Intelligence Group (GTIG) said it had identified exploit_agent, a multi-agent framework that automated vulnerability research and exploitation.The findings mark a shift from hackers using AI for individual tasks to deploying it across automated attack workflows.What are hackers doing?As per the report, this AI system can continuously scan software updates of widely used applications. From there, the AI can compare code before and after patches and identify fixed vulnerabilities. Once it has identified what was fixed in the new update, the AI – all by itself – can then plan targeted attacks at organisations who may not have updated to the new version. GTIG found exploit_agent on an exposed staging server, where it was connected to several frontier and locally hosted LLMs. The framework, as per Google, continuously analysed popular web application software by comparing code released before and after patch announcements. After identifying the fixed flaws, the system was able to plan and execute attacks on older versions of the software.Google said attackers may find it more efficient to use LLMs against patched vulnerabilities, known as n-days, than to search for previously unknown zero-day flaws. An announced patch gives attackers a limited set of material to analyse, including different versions of source code, release notes, security research and proof-of-concept code. The report said that this server also contained data taken from victims and scripts for internal reconnaissance, brute-force attacks, installing webshells and removing webshells deployed by rival attackers. Victim telemetry indicated that the operator was particularly interested in French websites associated with the media, entertainment and education sectors. A comment written in Russian in the framework’s configuration file instructed the operator to enter LLM API keys “right here, not in the chat”.The report cited two vulnerabilities in PaperCut print management servers that were exploited as zero-days in late August 2026. Security researchers and Huntress generated working exploits using LLMs within a day of disclosure. Four days after disclosure, GreyNoise identified a campaign that was believed to have used LLMs to build and test exploits before targeting hundreds of organisations. In one case, an attacker progressed from an initial foothold to domain administrator within minutes.AI is making hackers fasterAccording to the company, hackers hadn’t made a breakthrough when it came to exploits. Rather, AI tools made attacks quicker. “For the most part, attackers aren’t using new groundbreaking methods to break into systems — they are still going after known, unfixed issues, just with much greater speed and scale,” said John Hultquist, chief analyst at GTIG.Hultquist claimed that it was time for companies to use AI for cyberdefense to be able to stop such attacks. “Instead of relying on manual effort to keep up, we will have to embrace agentic defence. Or we risk falling behind,” he added.The company insisted that machine-speed threats could not be countered at human speed. Google is using its AI models and working with defenders to identify and fix flaws before a hacker may exploit them.Google’s report comes at a time when there is growing concern over cybersecurity. In recent months, hackers have increased attacks with the help of AI. One such case was the $6.3 billion Binance breach where hackers exploited an internal AI tool of the crypto platform to gain wider access to its systems.Last month, Microsoft claimed that a China-linked hacker group used AI framework, ARTEX, to launch automated cyberattacks against hundreds of organisations worldwide. Previously, CERT-In also warned that hackers were now using AI tools to target organisations.- Ends
Google says hackers are using AI agents to weaponise security patches without human intervention
Full Article
Original Source
Read the full article at Indiatoday →KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.