Microsoft (MS) has begun in earnest to capture the AI security market by unveiling its first cybersecurity-specialized artificial intelligence (AI) model and security automation platform.

As generative AI competition spreads beyond general-purpose chatbots into specialized cybersecurity fields, technological competition with Anthropic, Google, and OpenAI is expected to intensify.

According to HNGN on the 27th (local time), Microsoft unveiled the cybersecurity-dedicated AI model 'MAI-Cyber-1-Flash' and the AI-based security platform 'Project Perception' at an event held in San Francisco, USA.

Through this announcement, the company presented a strategy to cultivate vulnerability detection and automated response using AI as a new growth driver.

▲ Unveiling First Security-Specialized AI Model...Emphasizing Vulnerability Detection Performance

Microsoft's MAI-Cyber-1-Flash is an AI model specialized in identifying security vulnerabilities in complex software code.

The company announced that the model achieved a score of 96% on the 'CyberGym' benchmark, which is used as a standard for evaluating AI cybersecurity performance.

This is reportedly 12 percentage points higher performance than Anthropic's next-generation model 'Mythos'.

This model will be applied to the existing software vulnerability detection and remediation framework 'MDASH' to support vulnerability identification and security measure automation.

The market responded positively. Following the announcement, Microsoft's stock price rose approximately 3%.

▲ AI Agents Automate Security Tasks

Microsoft unveiled Project Perception, a security automation platform, alongside its AI model.

This platform is designed based on a structure where multiple AI agents collaborate to automatically perform continuous security monitoring, system patching, attack simulation, incident investigation, and vulnerability remediation.

The company described this as a 'machine-speed adaptive defense system' and emphasized it as technology that can change the paradigm of corporate security operations.

Dave Weston, principal engineer overseeing Project Perception, explained that AI can handle in minutes the tasks that application security experts and remediation engineers previously performed over hours.

The core competitive advantage highlighted includes the ability to automate most processes from vulnerability discovery through prioritization, detection, security configuration improvement, to code correction.

▲ AI Security Competition Expanding from General-Purpose AI to Specialized AI

This announcement was evaluated as an example showing that the stage of AI competition is rapidly shifting from general-purpose generative AI to industry-specific specialized AI.

Until now, AI companies have focused on conversational chatbots and productivity tools, but recently the development of specialized AI optimized for specific fields such as healthcare, finance, and security is emerging as a new competitive axis.

Microsoft's direct performance comparison with Anthropic, Google, and OpenAI models also demonstrates that it recognizes the cybersecurity market as an independent AI competition field.

Industry analysts suggest that specialized AI trained for specific fields could show higher efficiency in actual business operations compared to general-purpose large language models.

Microsoft Logo(Photo : [Reuters/Yonhapnews provided])

▲ Attackers Also Utilize AI...Security Competition a 'Double-Edged Sword'

The advancement of AI technology presents new challenges in that it is being utilized simultaneously in both defensive and offensive aspects.

Recently, hackers are increasingly using generative AI to discover software vulnerabilities more rapidly and automate attacks. Consequently, concerns are mounting that response speed is gradually reaching limits with only traditional manpower-focused security systems.

It is analyzed that Microsoft determined that establishing an automated security system where AI confronts AI is essential to respond to these environmental changes.

▲ Corporate Security Organization Changes Also Drawing Attention

Should Project Perception become commercialized, significant changes in corporate security operations are anticipated.

As AI performs repetitive analysis and response tasks, security personnel may be reorganized to focus on more complex threat analysis and strategy development.

Conversely, as automation expands in security operations centers (SOCs), changes in personnel demand may emerge in some simple operational areas.

▲ Preview Launch in November...Real Competition Begins

Microsoft announced plans to provide MAI-Cyber-1-Flash and Project Perception in preview form to enterprise customers starting November 3rd.

Since Anthropic, Google, and OpenAI are also developing cybersecurity-specialized AI technologies, competition in the specialized AI-based security market is expected to intensify further.