In an attempt to incline the battlefield of cybersecurity in favor of the defenders, Google has introduced sect-gemini V1, a new experimental AI model designed to help security equipment to identify threats, analyze incidents and understand the vulnerabilities faster and more precise than before.
Announced by the company's cybersecurity research leaders, Elie Burzstein and Marianna Tishchenko, Sec-Gemini V1 is the latest incorporation into the growing family of Google tools with Gemini, but this time, it focuses with laser on cybersecurity.
The growing cyber threat, and why Google's AI pushes matters
Cyber attacks are increasingly frequent, sophisticated and attacked. From ransomware to piracy sponsored by the State, defenders are overwhelmed. Add to that the increase in remote work, cloud systems and open source software, and the threat panorama becomes even more complicated.
Cybersecurity has always been an unfair fight. Attackers only need to find a weak point, while defenders must protect all possible entry points. Google's response is to develop an AI that acts as a force multiplier, helping human analysts work more intelligent. It is a Oneversus-Lall game, and Google believes that AI can help level the playing field.
What makes Sec-Gemini v1 different?
What distinguishes Sect-Gemini V1 is its access to real-time cybernetic safety data from sources of trust such as Google Threat Intelligence (GTI), Mandiant's attack reports and the open source vulnerabilities database (OSV). This leaves it:
- Identify the root causes of faster security incidents.
- Identify the threat actors (such as the group of salt tifones linked to Chinese) and their tactics.
- Analyze vulnerabilities in context, explaining not only what is broken but how computer pirates could exploit it.
Google states that the model has already shown strong results in internal tests, surpassing other models of the leaders, including OpenAi and Claude de Anthrope, at key security points. At the CTI-MCQ reference point, which measures how well the intelligence of threats understands, Sec-Gemini obtained a more 11% score higher. It also exceeded the rivals by 10.5% in the Ci-Root Mapeo test.
The largest IA security career
Google is not alone to boost safety with AI; Microsoft's backward co -pilot (Openai driven) and Amazon's guard are also betting on AI to automate the defenses. Even so, the deep integration of Google data and the performance of the reference points could give sec-gemini v1 an advantage, at least for now.
Google opens the doors, but only a little
IA security tools have had a mixed success. Some worry, they are only elegant assistants that still require human supervision. But Google insists that sec-gemini v1 is different. Not only summarizes threats, but it explains them in a way that accelerates decision making.
For now, it is only available for research, not for commercial use. But if it is up to the exaggeration, it could mark a turning point in how defenders remain up to date with computer pirates in a world loaded with AI.
Interested in trying Sec-Gemini V1? Google is taking requests through this form.
