seolounge
Perplexity AI has integrated the new Chinese DeepSeek AI model into their offerings, allowing their Pro level users to use DeepSeek for their Perplexity AI research. Some in the public reacted negatively to the news.
Perplexity AI
Perplexity AI is San Francisco based AI search engine that offers a different way to search for information by leveraging web content and large language models. There is also a Pro Search version that allows unlimited file uploads, can generate images and offers a choice between multiple popular AI models like OpenAI o1 and Anthropic’s Claude 3.5.
Now it’s offering DeepSeek R1 as one of the available choices for Pro Users. The announcement was met with misconceptions about what was being offered, including unfounded accusations that Perplexity DeepSeek data would be accessible to the Chinese communist government and that the search results would be censored.
Aravind Srinivas, Cofounder and CEO of Perplexity, commented on LinkedIn about the controversy:
“All DeepSeek usage in Perplexity is through models hosted in data centers in the USA and Europe. DeepSeek is *open-source*. None of your data goes to China.”
The CEO also took to X (formerly Twitter) to reassure users that the model they are using is not censored, posting a screenshot of an uncensored response to demonstrate that the version of DeepSeek R1 in use at Perplexity is not censored
Screenshot Of Uncensored Perplexity AI DeepSeek R1
Is DeepSeek Self-Hosted Censored?
Anyone can download the DeepSeek AI model and use it locally but the model as-is will be censored since it’s only good as the data it was trained on. The Register downloaded and tested multiple models of DeepSeek and concluded that it is indeed censored:
“Is it censored?
Oh yeah. It is. Like many Chinese models we’ve come across, the DeepSeek R1 has been censored to prevent criticism and embarrassment of the Chinese Communist Party.Ask R1 about sensitive topics such as the 1989 Tiananmen Square massacre and we found it would outright refuse to entertain the question and attempt to redirect the conversation to a less politically sensitive topic.
…Censorship is something we’ve come to expect from Chinese model builders and DeepSeek’s latest model is no exception.”
However, as Perplexity AI’s CEO Aravind Srinivas showed, the model can be uncensored. Contrary to some commenters on the LinkedIn discussion, a self-hosted model does not phone home back to China, everything is contained within the local environment.
Featured Image by Shutterstock/gguy