Google AI Threat Defense: Google has launched a new cybersecurity model. It's considered a rival to Cloud Mythos and Daybreak. However, its working method is slightly different.

Google AI Threat Defense: Google launches cybersecurity model

Google AI Threat Defense: American AI company Anthropic launched the Claude Mythos model last month. This model is so powerful that it can hack any software. In view of this danger, the company has not rolled it out to the public. To compete with it, OpenAI launched the Daybreak model a few days ago, and now Google has also jumped into this race. Google has introduced a new cybersecurity model, Google AI Threat Defense. Google says that this tool continuously monitors AI-driven threats and stops them before they have any effect.

Software flaws can be fixed immediately.

Google explained in a blog post that the tool allows companies to identify potential cyberattack patterns, prioritize the most dangerous threats, and quickly fix any vulnerabilities. This way, software vulnerabilities can be exploited.

Google's new model is different from Claude Mythos and Daybreak.

Google explained that a major problem with AI cybersecurity systems is that they constantly send thousands of AI-generated alerts to security teams. Cloud Mythos and Daybreak continuously detect vulnerabilities and send alerts, but not every vulnerability poses a significant threat. Google is now working to address this issue. This platform will integrate code scanning with its cloud-security platform to check whether a vulnerability can be exploited via the internet or live network configuration. If the platform finds a major vulnerability that won't impact the rest of the system, it will downgrade it. This will give developers and security teams enough time to work on active threats.

Google's model will work differently in this case, too.

A major difference between these platforms is bug fixes. OpenAI's system suggests patches based on bugs, but Google's platform will work differently. This platform will have AI agents, under human supervision, rewrite legacy code and convert it into modern programming languages. Furthermore, this platform will test whether patches released to fix a bug will work before deployment.