SKT enlists gated Claude Mythos AI to shield core infrastructure

South Korea-based operator SK Telecom has entered into a strategic partnership with Anthropic as part of the new Project Glasswing initiative to enhance the security of its core digital infrastructure. The telecommunications provider is among a select group of approximately 150 global organisations, including technology giants Samsung and SK Hynix, tasked with implementing advanced artificial intelligence models to safeguard critical systems. Through this collaboration, the operator will utilise a gated version of the Claude AI model, specifically tuned for identifying and mitigating complex security risks within large-scale codebases.

The primary objective of the Project Glasswing programme is to deploy highly sophisticated AI tools to scan software architectures for deep-seated vulnerabilities that traditional security measures might overlook. By integrating this intelligence into its operations, SK Telecom aims to create an automated layer of defence capable of detecting external threats and internal flaws in real-time. This move comes as telecommunications networks increasingly transition toward software-defined models, which necessitates more robust and proactive cybersecurity frameworks to maintain service continuity.

Under the terms of the engagement, the AI model will be utilised to conduct comprehensive audits of the software that powers the operator’s mobile networks and data centres. The gated nature of the Claude model ensures that sensitive proprietary data remains within secure parameters while leveraging the massive processing power of generative AI. This approach allows for the rapid identification of potential exploit vectors, significantly reducing the window of opportunity for cybercriminals to compromise critical national infrastructure or customer data.

This development follows broader industry trends where South Korean technology companies are aggressively integrating artificial intelligence into their fundamental management structures. SK Telecom has been particularly active in the AI space, seeking to transform from a traditional connectivity provider into a global AI company. By participating in this high-level security initiative alongside its domestic hardware partners, the mobile operator is strengthening the domestic technology ecosystem against evolving global cyber threats.

The collaboration signifies a shift toward using artificial intelligence not just for consumer-facing services, but as a foundational security tool for the underlying telecommunications fabric. The technical team at the operator will work closely with system architects to ensure the AI-driven scans are integrated into the continuous development and deployment cycles. This integration is expected to streamline the patching process and ensure that new software features meet stringent security standards prior to being launched on the live network.

Looking ahead, SK Telecom intends to refine its security protocols as the Claude-integrated systems provide more granular insights into network vulnerabilities. The success of this implementation will likely serve as a blueprint for how other international operators adopt gated AI models for infrastructure protection. As the rollout progresses, the company plans to further automate its threat response mechanisms to maintain a high level of resilience across its South Korean and international operations.