Google’s generative AI efforts in cybersecurity began in earnest in 2017 with the release of Transformer’s new neural network architecture. The company notes that recurrent neural networks are now at the heart of its approach to language modeling.
Fast-forward to RSA 2023, and Google Cloud announced that it will use generative AI and a security large-scale language model (LLM), Sec-PaLM, to significantly enhance security.
But why now? Steph Hay, Director of User Experience at Google Cloud, says LLM has reached a critical mass of being able to contextualize information in ways it couldn’t before.
“We now have truly generative AI. It’s not rules-based, it’s simply call-and-response,” said Steph Hay, Director of User Experience. Information security.
The company announced Google Cloud Security AI Workbench, the industry’s first scalable platform powered by Sec-PaLM.
Sec-PaLM incorporates Mandiant’s leading intelligence on vulnerabilities, malware, threat indicators, and behavioral attacker profiles.
Hay agreed that Google’s acquisition of Mandiant, which will be completed in 2022, is a key component for generative AI. This allows the company to have industry-leading threat intelligence to train his LLM.
Additionally, Security AI Workbench is built on Google Cloud’s Vertex AI infrastructure, where customers manage their data with enterprise-level features such as data isolation, data protection, sovereignty, and compliance support, according to Google. Cloud explains.
The company is also introducing Sec-PaLM into VirusTotal Code Insight to analyze and describe the behavior of potentially malicious scripts, allowing it to better detect which scripts are threats.
Mandiant Breach Analytics for Chronicle also uses Sec-PaLM to contextualize these key findings for immediate action.
This contextualization is one reason Hay believes the age of generative AI in cybersecurity has arrived.
Amid all the excitement new to the company’s new Sec-PaLM, Hay warned that the coming months will be “critical” for the platform.
Google Cloud believes generative AI will revolutionize security. Because it helps fight threats, reduce effort, and elevate your workforce beyond cybersecurity.
The company is looking for partners to work with Security AI Workbench, but it was clear that this did not mean that the partner’s data would be used to train Sec-PaLM. An opt-out option will be available.
Editorial image credit: Monticello / Shutterstock
