Google Cloud has announced a new unified security platform for enterprises, delivering new interoperability and AI capabilities for cyber teams and reducing time spent investigating threats in cloud environments.
Google Unified Security combines products, telemetry, and context from Google Threat Intelligence, Mandiant, Google Security Operations, and Google Security Command Center, to provide organizations with an intelligent overview of threats.
Referred to by executives as ‘GUS’, the offering is a searchable data fabric for enterprise vulnerability, tying together the data and capabilities of the above platforms for better security visibility.
Gemini AI sits at the core of Google Unified Security, as a key enabler for decreasing time to detection for security teams and reducing manual workload. Google Cloud’s internal AI models will power features such as two new security agents that will work alongside human cyber workers on proactive, low-level analysis and producing reports.
A new alert triage agent, available within Google Security Operations, will continuously analyze enterprise environments for alerts and flag any activity it deems worthy of human response.
The second, a new malware analysis agent contained within Google Threat Intelligence, will proactively scan the code for all uploaded files for malicious code. Matching this against known malware examples, it will then provide a detailed summary for security teams.
Both agents will become available to certain customers via preview in Q2 2025, with general availability at a later date.
In a demo, Payal Chakravarty, director of Product Management at Google Cloud, showed a data loss prevention (DLP) alert in Google Chrome Enterprise flagging an incident in which developer has inadvertently copied sensitive data into public LLMs through a Chrome extension.
Responding to the alert, a Gemini agent is then able to investigate the situation and confirm the leak to a high degree of confidence, making detailed logs throughout and quarantining the Chrome extension involved. The agent can then go further by updating the company’s policy on the extension to prevent any further data leaks.
Similarly, Chakravarty explained that the agent could continue to investigate and discover other vulnerabilities such as the developer accidentally exposing a virtual machine (VM) to the public internet. It could then correlate any suspicious traffic to the VM with known signatures from active threat actors using data from Google Threat Intelligence.
Brian Roddy, VP of cloud security at Google Cloud, explained that GUS is intended to meet the needs of businesses in an era of increasing attacks by state-backed threat actors and criminal cyber gangs.
“Enterprise infrastructure continues to grow in size and complexity, expanding the attack surface and making defenders’ jobs increasingly difficult and there’s separate, disconnected security tools resulting in a fragmented data situation, without relevant context, leaving organizations vulnerable and reactive in the face of escalating threats.”
“So security teams tend to operate at silos, slowed by toils of workflows, making it hard to accurately assess and improve the organization’s overall risk profile.
The announcement was made live at Google Cloud Next 2025, the company’s annual conference held in Las Vegas.
Managing AI risk and common vulnerabilities
Google Cloud has recognized that along with the benefits AI presents businesses, improper use or implementation of the technology can also expose organizations to unwanted risks.
In March, Google Cloud announced AI Protection, through which leaders can take stock of the AI models and tools they use across their environment and secure data used for AI.
Model Armor, a feature integrated within Vertex AI, allows leaders to apply specific controls across AI inputs and outputs across their cloud environment, across a wide range of models.
Now Google Cloud has announced new Data Security Posture Management features, which flags and labels sensitive data, puts it under suitable compliance controls, and monitors how it is used for AI directly within the Google Cloud AI portfolio and analytics tools such as BigQuery.
Alongside this, a newly-launched Compliance Manager will work to give users better understanding of their data compliance and help internal teams produce audits to keep on top of reporting requirements. Both features will enter preview in June.
In response to an ITPro question on how Google Cloud customers are tackling issues such as shadow AI, Roddy stated that early adopters are already finding AI Protection useful for detecting unauthorized AI model use within their workplace.
“So what they want to have is the ability to control, what are really the models that are approved and tested and which ones aren’t,” Roddy said.
“And in some cases they’re willing to let users use more advanced models for experimental purposes. So what this is doing is just giving them a lot more control about understanding what’s inside their environment and to be able to make sure that the right models are being used for the right use cases.”
Google Chrome Enterprise Premium is also set to receive new privacy controls, including data masking to automatically hide sensitive information and phishing protection to prevent man in the middle (MiTM) attacks in which malicious sites are disguised as legitimate web portals.
This is in addition to the rollout of existing Google Chrome Enterprise Premium features, such as copy paste prevention and URL filtering, across Android.
