[DigitalToday reporter Yoonseo Lee] OpenAI CEO Sam Altman (샘 알트먼) said artificial intelligence is quickly exposing security vulnerabilities built up in existing systems and companies need to change how they respond.

Tech outlet TechRadar reported on Saturday that Altman, in an on-stage discussion with Oracle co-CEO Clay Magouyrk (클레이 머고이크), pointed to the limits of the traditional approach of fixing individual bugs one by one.

Magouyrk said AI is helping find security problems accumulated over the past 50 years. He said AI is not creating new vulnerabilities but identifying issues hidden in existing systems.

Oracle’s latest quarterly security patch update addressed about 3,000 issues, and most of them were found through AI, the report said. As AI speeds up vulnerability detection, companies’ patching work is also increasing quickly.

Oracle also introduced Oracle AI Security Mesh, built in cooperation with OpenAI, to support companies’ vulnerability detection and response. It is designed to provide supporting material, including steps to reproduce vulnerabilities and test results, so security teams can filter false positives and respond first to higher-risk issues. It also supports investigating vulnerabilities and reviewing remediation plans before an actual attack occurs.

AI’s advances could also work to attackers’ advantage. Johannes Landgraf (요하네스 란트그라프), OpenAI’s head of cyber, said advanced AI offers new opportunities for organisations to detect and respond to sophisticated security threats. At the same time, concerns are growing that attackers can also use AI to find and exploit vulnerabilities faster.

Altman said, "It can be too hard to patch every bug across the entire supply chain one by one," and called for new ways to fix problems. The remark is seen as indicating the need to shift security strategy from individual bug fixes to constant monitoring. As AI use expands, the role of security teams and how organisations operate could also change.

Companies’ task is to quickly deploy fixes in step with the pace of vulnerability discovery. Even if AI improves detection, delayed action can leave organisations exposed to attacks, making it increasingly important to reduce the gap between discovery and response.