Skip to main content
AI Learning Resources

Quick Overview

This video is a case-study interview produced by OpenAI featuring Gregor Stewart, Chief AI Officer at SentinelOne. It addresses the practical application of generative AI models in cybersecurity research, specifically focusing on malware analysis and threat intelligence workflows.

Key Points

  • 1.Security teams have historically been constrained by the fact that researchers can reverse engineer only a very small number of software samples.
  • 2.Artificial intelligence has reached human-level capability in reverse engineering malware software.
  • 3.AI reduces reverse engineering tasks that took a top-flight researcher four to six weeks down to roughly eight hours.
  • 4.Rapid reverse engineering allows defenders to identify shared elements across software and map relationships between previously unrelated malware.
  • 5.OpenAI is shifting the advantage to defenders by providing expert-level cybersecurity capability to everyone on the development team.

Summary

Gregor Stewart, Chief AI Officer at SentinelOne, opens by questioning how much of the world of malware can realistically be mapped, given that analysts are currently able to reverse engineer only very small numbers of software samples. He explains that artificial intelligence has now reached human-level capability in reverse engineering. To demonstrate the impact, he states that a task requiring four to six weeks for a top-flight researcher can now be completed at human-level output in roughly eight hours.

Stewart points out that this accelerated speed fundamentally changes threat intelligence. When an analyst identifies an element inside a piece of software that was previously observed elsewhere, knowing who built it allows defenders to establish a relationship between two pieces of malware that were previously thought to be unrelated. He notes that this level of intelligence was something organizations were never able to afford before due to time and resource constraints.

Finally, Stewart describes the broader strategic consequence for cybersecurity teams. He explains that OpenAI is shifting the advantage toward defenders by delivering expert-level cyber capability directly to every member of the development team, expanding reverse engineering power beyond a small group of specialized researchers.

Accelerating Reverse Engineering of Malware

Cybersecurity researchers face significant scaling limits because reverse engineering software traditionally requires extensive manual effort from top specialists. With AI achieving human-level capability, complex reverse engineering tasks that previously required four to six weeks can now be completed in approximately eight hours.

Mapping Relationships Across Unrelated Malware

Automating reverse engineering at scale allows analysts to identify shared code elements across different software samples. Recognizing author signatures and code reuse enables security teams to uncover previously unseen relationships between different malware variants, creating threat intelligence that was formerly unaffordable.

Shifting the Cybersecurity Advantage to Defenders

OpenAI shifts the balance in cybersecurity toward defenders by broadening access to advanced analytical capabilities. Instead of keeping expert-level reverse engineering confined to a few specialized researchers, these capabilities are made available to everyone across the development team.

The Bottom Line

The video establishes that AI models can match human-level reverse engineering performance, reducing malware analysis timelines from weeks to hours and enabling defenders to uncover relationships between disparate malware samples. It concludes that deploying expert-level capabilities across entire development teams helps shift the strategic cybersecurity advantage back to defenders. However, the short discussion leaves unresolved what specific technical models or validation workflows SentinelOne uses to verify the accuracy of AI-generated reverse engineering outputs.

FAQ

What is SentinelOne and what does SentinelOne do with OpenAI in cybersecurity research?

SentinelOne is a cybersecurity company whose Chief AI Officer, Gregor Stewart, discusses using OpenAI technology to achieve human-level reverse engineering of malware software, accelerating analysis from weeks to hours.

What limitation do cybersecurity researchers currently face when attempting to map the world of malware without AI?

Researchers are currently able to reverse engineer only very small numbers of software samples because manual analysis is too slow and resource-intensive to scale.

How much time does OpenAI save researchers when reverse engineering malware software compared to human analysts?

Analysis that typically took a top-flight researcher four to six weeks can be completed at human-level output in approximately eight hours.

How does AI reverse engineering help cybersecurity defenders identify relationships between different pieces of malware software?

By detecting shared code elements previously identified in other software and attributing authorship, AI helps researchers connect pieces of malware that previously appeared unrelated.

How is OpenAI shifting the advantage to cybersecurity defenders across software development teams according to SentinelOne?

OpenAI shifts the advantage to defenders by delivering expert-level cybersecurity and reverse engineering capabilities to everyone on the development team.

Worth watching for

Cybersecurity professionals, threat intelligence analysts, and software development leaders interested in how AI accelerates reverse engineering and malware analysis.

  • cybersecurity
  • malware
  • reverse-engineering
  • sentinelone
  • openai
Watch on YouTubeMore videos from OpenAI