The artificial intelligence open-source large language models (LLMs) application developed by a Chinese AI company DeepSeek is as good as ChatGPT, but there are some concerns, Dr. Manjeet Rege, software engineering professor and director of the Center for Applied Artificial Intelligence at the University of St. Thomas, said during an interview on the Adam and Jordana WCCO Radio show.
“If you download it on your phone, you are giving away all of your data,” he said.
From the interview:
One has to also consider the security risk here in terms of national security. So, depending on what aspect of DeepSeek used, we could be giving our data to the Chinese government. So, if you download DeepSeek on your phones or if you use the web interface, DeepSeek’s privacy policy states that all of that information – chat interactions, our keystrokes, our IP addresses – they could be utilizing for further refining their model and that information is being stored in China. Now if you download the open source model locally on your computer, then that information is not being sent to China.
From a related interview Rege did with KSTP-TV:
The policy also states it keeps the data “as long as necessary” to have the app function and that it stores information “in secure servers located in the People’s Republic of China.”
“Scrutiny from experts should somewhat alleviate concerns about privacy and security as well,” said Manjeet Rege, director of the University of St. Thomas’ Center for Applied Artificial Intelligence.
