Use 'DeepSeek' Safely on KakaoTalk 'Luiten' Channel

Safe Service with Separate Cloud Model Deployment
"User Input Data Will Not Be Leaked"

AI service platform Lutton Technologies announced on the 4th that it will be the first in Korea to launch the 'Safe Service' of China's DeepSeek R1 AI model on the KakaoTalk Lutton channel.

The latest R1 model from DeepSeek, released on the 20th of last month, has attracted significant attention by demonstrating performance equal to or higher than OpenAI's existing models, ranking number one in app downloads worldwide, including in the US and Korea. However, due to controversies over personal information security vulnerabilities, unauthorized data collection, and suspicions of model training misuse, some countries have blocked downloads or restricted public use.

Lutton dispelled these concerns by providing the service in a safe manner, hosting the model on a separate cloud rather than using DeepSeek's own service. A Lutton representative stated, "The DeepSeek Safe Service provided by Lutton operates the model on a cloud physically separated from the model creator, so user input data is not leaked to any specific country and is not used for the creator’s model training."

The DeepSeek Safe Service offered by Lutton is available for free and unlimited use on the KakaoTalk Lutton channel. By searching for and adding the Lutton channel as a friend on KakaoTalk, users can chat with DeepSeek in the conversation window. Lutton plans to first provide the DeepSeek model service on the KakaoTalk channel, then analyze actual performance and user feedback to decide whether to expand the service to its own web and app platforms in the future.

Se-young Lee, CEO of Lutton, stated, "Even amid the upcoming fierce competition among new AI models following DeepSeek, we will continue to strive to ensure that everyone can enjoy AI in the fastest, most convenient, and safest way."

This content was produced with the assistance of AI translation services.