Anthropic has opened a new office in Seoul, marking an expansion of its presence in South Korea.

The company also announced a memorandum of understanding with Korea’s Ministry of Science and ICT. The stated aim is to support the “safe and responsible” adoption of AI across the public sector.

What’s inside the agreement is less about broad public messaging and more about practical risk work. Anthropic said the MOU includes collaboration on AI safety and cybersecurity.

That includes evaluating model safety in Korean with the Korea AI Safety Institute. Anthropic also said the two sides will share information about AI-enabled cyber threats, which suggests an effort to connect AI safety testing with real-world security concerns.

This matters for how AI ends up being used inside government systems. Safety evaluations that are tailored to local language and context can be a key step, especially when public services depend on AI tools. Cybersecurity coordination is also important because AI deployments can create new pathways for attackers, even when the goal is safer, more reliable service delivery.

For now, the office opening and the MOU give a clear signal that Anthropic is moving to build relationships with local institutions, not just sell or deploy technology remotely. The company didn’t provide additional details beyond the MOU’s focus areas, but the scope—model safety work in Korean and sharing cyber threat information—points to an implementation-minded approach.

Source: Anthropic