Google introduces generative AI-based security features

Google has launched a Security Artificial Intelligence Workbench based on a large language model Sec-PaLM. Sec-PaLM is an offshoot of Google’s Pathways Language Model (PaLM) launched last year but is fine-tuned for security use cases. The announcement was made by Google at the RSA conference held yesterday.

The Security AI Workbench from Google Cloud is an industry-extensible cybersecurity suite that addresses threat overload, talent gaps, and complex tools. As per Google, the AI Workbench also features partner plug-in integrations to offer Google Cloud’s customers capabilities like threat intelligence, workflow, and other critical security functionalities. Accenture has already begun using the Security AI Workbench.

The security platform is built on Google Cloud’s Vertex AI infrastructure. This enables customers to control their data through data isolation, data protection, sovereignty, and compliance support.

The underlying model beneath Security AI Workbench – Sec-PaLM – is based on ‘years of foundational AI research by Google and DeepMind’, the company said in its blog. Further adding that the work includes new efforts to expand partner ecosystem and provide businesses with security at all levels of the cybersecurity stack. “We have only just begun to realise the power of applying generative AI to security, and we look forward to continuing to leverage this expertise for our customers and drive advancements across the security community,” the blog said. 

As part of the Security AI Workbench, Google Cloud’s customers will get access to VirusTotal Code Insight, which uses Sec-PaLM to analyse and explain the behaviour of potentially malicious scripts. Google Cloud is also offering Madiant Breach Analytics for Chronicle to alert about active breaches; this feature is based on cybersecurity firm Madiant’s tech which was acquired by Google in September 2022.

Other features launched include Assured OSS that will improve open source software (OSS) vulnerability management, and Madiant Threat Intelligence AI to quickly find, summarise and act on threats.

Notably, Microsoft also introduced Security Copilot, based on GPT-4, in March. It aims to summarise threat intelligence and augment the work of security professionals through an ‘easy-to-use AI assistant’.