The Deepseek apps have been temporarily blocked in South Korea after a government investigation discovered that its artificial intelligence model had sent personal data to China’s Bytedonce firm. Previously, the South Korean intelligence agency had accused the Chinese firm of “excessive” collection of personal data and use for training your language model.
Deepseek surprised the technology and investor industry for its ability to offer functions of benefits similar to large western models, to a fraction of the cost. Deepseek’s progress suggested that Chinese engineers would have found a way to overcome chips export prohibitions, focusing on achieving greater efficiency with limited resources. In addition to “clicking” the AI bubble, the entry into the scene of this startup confirmed that China is willing to compete in AI.
Deepseek and privacy
Several countries have questioned the privacy of user data storage by Deepseek, which according to the company are collected in «Insurance Servers located in the People’s Republic of China«.
Italia He initiated an investigation into Depseek’s R1 model and filed a blockade to prevent him from processing user data in the country. For its part, AAustralia The assistant has prohibited in all government devices following the Council of Security Agencies, while the US legislators have also proposed a bill to prohibit the use of Deepseek on government devices due to concerns about user data safety.
Although, in response to prohibitions, the Chinese government has insisted that “It has never required and will never require companies or individuals to collect or store illegally”, The suspicions continue worldwide and the last veto has arrived from South Korea. “We discovered that Deepseek users’ data leaked to Bytedonce»the Personal Information Protection Commission of Korea said on Monday.
Deepseek is not currently available in South Korean digital stores of iOS and Android, although those who already had it can use the service, as well as web. “The commission is in the investigation stage to determine if Depseek raises any damage”the regulator explained. “A large -scale prohibition could be controversial before its responsibility is officially determined”he said.
The regulator expects to complete its research faster than that carried out on Openai, Google and Microsoft last year, which took five months. And is that, On privacy field, no company is saved. And less with the insatiable AI models that need a countless amount of data to feed.
