OX Security is shifting security as far left as it can go with the launch of VibeSec, which it says can stop insecure AI-generated code before the code even gets generated.
It does this by embedding dynamic security context into the coding model so that it doesn’t suggest code that contains security issues.
“VibeSec doesn’t just accelerate security – it fundamentally changes how security operates. For the first time, security moves faster than vulnerabilities,” said Neatsun Ziv, co-founder and CEO, at OX Security.
According to recent data from BaxBench, 62% of AI-generated code contains flawed or vulnerable code, and another study found a 37.6% increase in critical vulnerabilities after an LLM does five iterations on code. OX Security believes this confirms that the more AI is applied to coding, the more insecure it becomes, and why the company created VibeSec in response to the vibe coding movement.
VibeSec is built on top of the company’s security intelligence engine OX Mind. OX Mind includes an AI Data Lake that provides alignment between security and company-specific code, cloud infrastructure, APIs, and runtime environments.
OX Mind also analyzes the company’s infrastructure, architecture, and codebase to provide tailored preventative actions and prioritizations. Finally, it integrates security policies into development workflows to ensure compliance at every stage of the process.
“The old plugin model was built for human typing speed. The new reality is AI-driven code generation at machine speed, and that demands an equally new security model – one that doesn’t just nudge developers, but aligns centrally with how code is actually generated now. This new model must bring in dynamic real-time context from APIs, code, runtime, and cloud to guide how new, secure code is created and validated,” Ziv said.
The post VibeSec embeds security analysis into AI coding models to prevent generation of insecure code appeared first on SD Times.
from SD Times https://ift.tt/34Awcu0
Comments
Post a Comment