Gregor Stewart, Chief AI Officer at SentinelOne, has presented the potential of using OpenAI models in malware analysis. New tools are capable of analyzing software at the level of a human expert, reducing the time required to examine samples from weeks to just a few hours. For the security industry, this means faster detection of links between attacks and the democratization of advanced defense methods for the entire engineering team.
Manual analysis of suspicious code has been one of the most time-consuming tasks in cybersecurity for years. As Gregor Stewart, Chief AI Officer at SentinelOne, points out, the application of modern artificial intelligence models is changing this process in a way that is tangible for technical teams.
Code analysis in hours instead of weeks
According to information provided by Stewart, artificial intelligence is now capable of mapping and analyzing malware at a level comparable to a human. Reverse engineering tasks that previously took analysts weeks of work are now performed by automated systems in just a few hours.
Identifying links between threats
Accelerating workflows is not the only aspect discussed by the SentinelOne expert. Equally important is the ability of algorithms to detect correlations between different pieces of software. The tools can compare samples and identify dependencies, which allows for linking threats that were previously considered entirely separate.
New capabilities for development teams
According to Stewart, the use of technologies developed by OpenAI shifts the balance of power in cyberspace in favor of defenders. Advanced analytical capabilities are no longer the domain of only a narrow group of specialized malware researchers — such tools can now be used by the entire team of engineers and developers responsible for infrastructure security.
Comments