检查点如何与"黎明"一起加速脆弱性研究

中文版正在随流水线自动整理，稍后刷新即可；可先阅读下方英文原文。

本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。

How Check Point accelerates vulnerability research with Daybreak

With OpenAI Daybreak, Check Point speeds up the research behind new security protections, enabling its analysts to turn vulnerability intelligence into customer defenses faster.

Results

Hours to minutes

Vulnerability research with Daybreak

Check Point’s vulnerability research workflow used to take hours to complete. As part of the OpenAI Daybreak Defense Network, the company uses Daybreak to reduce that work to minutes, according to Jonathan Zanger, CTO of Check Point. Daybreak gives defenders access to OpenAI’s frontier cyber models that can analyze vulnerabilities, reason through attack paths, and identify security gaps in code and configurations.

Check Point helps secure over 100,000 organizations through its AI, workspace, cloud, and network security products. Check Point Research brings vulnerability research expertise to the company’s prevention-first approach, investigating emerging vulnerabilities and translating findings into protections for customers. Daybreak accelerates that research, shortening the path from understanding a vulnerability to developing a defense.

“OpenAI is putting frontier cybersecurity intelligence in the hands of defenders and enabling companies like Check Point to turn this intelligence into threat prevention.”

“We’re seeing more cyber threats, and those threats are becoming increasingly sophisticated,” Zanger says. For Check Point, keeping pace means quickly turning information about newly disclosed vulnerabilities into protections for customers. “When a new vulnerability emerges, our job is to make sure our customers are protected against it as fast as possible.”

That work depends on expert analysts who investigate vulnerabilities and translate their findings into security signatures or indicators of compromise that Check Point’s systems use to recognize threats. This research requires specialized expertise, and the team’s capacity was previously constrained by the number of analysts available to do it.

Check Point uses OpenAI Daybreak to automate more of this workflow. “We leverage Daybreak to speed up and automate this process, and we can now do it in minutes instead of hours,” Zanger says. The value comes from the models’ ability to reason about security problems.

“OpenAI Daybreak can understand vulnerabilities, analyze different attack patterns and scan the code to identify gaps.”

Zanger also points to advances in understanding complex code and configurations, and identifying how systems could be exposed to attack. Check Point now considers those capabilities mature enough to incorporate into its cybersecurity workflows. For vulnerability research, Daybreak supports the investigative work that informs the development of a signature or indicator of compromise, giving analysts a faster path from understanding a threat to developing protections.

Check Point is also working with OpenAI to develop its Agentic Network Security Orchestration solution using Daybreak. The approach Zanger describes would combine continuous monitoring of customer environments with clearly defined operating boundaries and human approval for consequential actions. Customers would set their security objectives, and the system would explain its findings, recommended actions, and associated risks so they could decide how to proceed.

With vulnerability research now taking minutes to complete, Check Point’s next priority is remediation. The company is looking to use Daybreak’s capabilities to make it easier for their security teams to address the gaps they identify.

“Combining OpenAI’s Daybreak with our existing security workflows unlocks a level of automation that can prevent threats in real time.”

For Check Point customers, faster research supports the development of protections against newly discovered vulnerabilities. With Daybreak, Check Point aims to extend that progress toward more autonomous defenses, with customers setting security objectives and approving consequential actions.

Auto-fetched and translated by AI Atlas. All rights belong to the original author.