Chinese startup DeepSeek AI has garnered attention this week due to its open-source language models, which face scrutiny over privacy and censorship issues linked to the Chinese government. Perplexity recently integrated DeepSeek's R1 model, allowing users some accessibility but with important limitations. While data can be safeguarded via US/EU servers, using the assistant app increases security risks because of its connection to Chinese servers. Testing revealed the model's capabilities but highlighted ongoing challenges surrounding enforced censorship, posing questions about its utility in sensitive topics.
"None of your data goes to China," Perplexity CEO Aravind Srinivas reiterated, emphasizing data safety while using DeepSeek's models.
Perplexity's Pro Search mode offers transparent reasoning, increasing access to the DeepSeek R1 model while preserving user data security in Western servers.
While DeepSeek’s models match top competitors, concerns about censorship and data security remain significant, especially with their ties to Chinese servers.
Testing of DeepSeek R1 revealed promising coding abilities, yet inherent censorship influenced by the Chinese government presents ongoing challenges.
Collection
[
|
...
]