At RSA Conference 2023, Google Cloud announced Google Cloud Security AI Workbench, a security platform powered by Sec-PaLM, a large language model designed particularly for cybersecurity use cases. Sec-PaLM modifies the existing PaLM model and combines the Google threat intelligence data with Mandiant’s frontline intelligence to find, neutralize, and coordinate response actions against hostile elements.
Eric Doerr, VP of Engineering at Google Cloud, during an interview with VentureBeat stated that with Sec-PaLM, it is possible to generate an auto-generated security policy, security control, or security configuration that goes with any infrastructure. Among the tools included as part of Google Cloud Security AI Workbench is VirusTotal Code Insight, released today in preview and allowing the user to assess the script for malicious activity. Another preview release in summer 2023 is Mandiant Breach Analytics for Chronicle, that sends user notification for potential breaches with Sec-PaLM to summarize and respond to the threats found.
The need for generative AI in defensive roles from security organizations has been observed as a market estimated to reach a value of $51.8 billion by 2028 according to MarketsandMarkets. SentinelOne also announced its own large language model security solution at RSAC today, which is powered by GPT-4 to speed up human-led threat-hunting investigations that can be automated with responses. Microsoft is also experimenting with defensive generative AI use cases with their AI assistant Security Copilot, which integrates their data with GPT-4 to filter and process the threat signals while also creating a summary of breach activity. Orca Security and Kubernetes security company ARMO offer integrations that use generative AI to automate SOC operations.
The data accumulated by Google Cloud across multiple services such as Mandiant threat intelligence, Chrome, Gmail, and YouTube makes the platform stand out. Doerr points out that customers will be able to directly use the LLM or plug their own data for refinement. With AI advancements in Google Cloud Security AI Workbench, the organization hopes to stay ahead and excel at the current market competition.
