Concerns are emerging that AI models such as Anthropic Mythos could be abused for cyberattacks. With hacking incidents linked to AI supply chains also occurring, expectations are rising that security firms' growth potential will increase.

That appears to be having a positive impact on the share prices of major security firms. Some see AI models as potentially replacing the role of existing security firms, but for now the dominant view appears to be that rising risks from the spread of AI will help security firms grow. CrowdStrike and Palo Alto Networks shares have each risen about 20 percent over the past month. BlackRock IGV, an index that tracks the software sector, rose only about 10 percent over the same period.

• Shares of major global security firms rise; some analysis says they are still undervalued • [Tech Insight] AI attack vs defence: defence will have the advantage in the long term

CrowdStrike introduced Project QuiltWorks, an alliance of companies that helps businesses quickly find and fix software vulnerabilities discovered by frontier AI models.

• CrowdStrike-led AI security threat response alliance 'Project QuiltWorks' gains attention

Anthropic, meanwhile, launched a public beta of Claude Security, a security tool that detects vulnerabilities in corporate codebases and generates patches. Claude Security analyses the entire codebase in a way similar to cybersecurity researchers. Instead of looking for known patterns, it tracks data flows, reads source code and analyses interactions between code components.

• Anthropic launches public beta of 'Claude Security' to detect software vulnerabilities and create patches

The following is a roundup of moves by companies in and outside South Korea related to AI.

OpenAI introduced Advanced Account Security (AAS), which users can use if they want. OpenAI said it was designed primarily for people in high-risk jobs, but anyone can sign up.

• OpenAI cooperates with Yubico to link security keys to ChatGPT accounts

Global security company Netskope warned of AI data leakage risks in its 2026 Cloud and Threat Report. It said it would strengthen its security strategy based on its AI-integrated security product, Netskope One AI Security.

• "Data leaks rise six-fold amid AI spread...need to redesign security strategy"

AI security specialist Aim Intelligence will cooperate with Mangoboost to target the next-generation AI infrastructure security market. The two companies will develop technology to boost AI infrastructure efficiency by combining an AI guardrail service with DPU-based server infrastructure. Winning i, a company that offers multi-biometric authentication and digital identity authentication solutions, has been invited to present at the Technology Innovation Forum 2026 event hosted by JPMorgan Chase in India.

Initech introduced an ICAM (Identity, Credential, and Access Management) platform targeting the zero trust security market. Privacy-enhancing technology firm DESILO will implement in commercial software and launch the GL Scheme (Gentry-Lee Scheme), which it bills as fifth-generation fully homomorphic encryption (FHE) technology. API governance specialist Softfreak said its system access control solution i-SeMOS obtained a security function confirmation certificate from South Korea's National Intelligence Service.

• Aim Intelligence partners with Mangoboost..."Developing hardware-based AI security" • Winning i to participate in JPMorgan Technology Innovation Forum...to unveil multimodal biometric platform • Initech launches zero trust-based ICAM solution...supporting response to N2SF adoption • DESILO launches commercial homomorphic encryption software based on 'GL Scheme'..."Private AI becomes reality" • Softfreak obtains National Intelligence Service security function confirmation certificate with system access control solution

Palo Alto Networks' Unit 42 research team developed Zealot, a proof-of-concept system to test whether AI can autonomously hack cloud environments.

• AI hacking agent Zealot succeeds in cloud infiltration and data theft without instructions

An expert also argued that combining open-source models can deliver vulnerability detection performance comparable to Anthropic Mythos. Ari Herbert-Voss (아리 허버트-보스), chief executive of AI security startup RunSybil, said this in a presentation at the Black Hat Asia conference held in Singapore.

• "Combining multiple open-source models enables Mythos-level vulnerability detection"

The government will push to distribute guidelines and provide consulting for small and medium-sized enterprises to respond to security threats from Anthropic's AI model Mythos.

• After warning on Anthropic 'Mythos', government says it will expand security support for SMEs