Perplexity AI has built-in the brand new Chinese language DeepSeek AI mannequin into their choices, permitting their Professional stage customers to make use of DeepSeek for his or her Perplexity AI analysis. Some within the public reacted negatively to the information.
Perplexity AI
Perplexity AI is San Francisco based mostly AI search engine that gives a distinct solution to seek for info by leveraging net content material and huge language fashions. There’s additionally a Professional Search model that enables limitless file uploads, can generate photos and presents a alternative between a number of widespread AI fashions like OpenAI o1 and Anthropic’s Claude 3.5.
Now it’s providing DeepSeek R1 as one of many obtainable selections for Professional Customers. The announcement was met with misconceptions about what was being provided, together with unfounded accusations that Perplexity DeepSeek information can be accessible to the Chinese language communist authorities and that the search outcomes can be censored.
Aravind Srinivas, Cofounder and CEO of Perplexity, commented on LinkedIn in regards to the controversy:
“All DeepSeek utilization in Perplexity is thru fashions hosted in information facilities within the USA and Europe. DeepSeek is *open-source*. None of your information goes to China.”
The CEO additionally took to X (previously Twitter) to reassure customers that the mannequin they’re utilizing isn’t censored, posting a screenshot of an uncensored response to reveal that the model of DeepSeek R1 in use at Perplexity isn’t censored
Screenshot Of Uncensored Perplexity AI DeepSeek R1
Is DeepSeek Self-Hosted Censored?
Anybody can obtain the DeepSeek AI mannequin and use it domestically however the mannequin as-is will probably be censored because it’s solely good as the info it was educated on. The Register downloaded and examined a number of fashions of DeepSeek and concluded that it’s certainly censored:
“Is it censored?
Oh yeah. It’s. Like many Chinese language fashions we’ve come throughout, the DeepSeek R1 has been censored to stop criticism and embarrassment of the Chinese language Communist Get together.Ask R1 about delicate matters such because the 1989 Tiananmen Sq. bloodbath and we discovered it could outright refuse to entertain the query and try and redirect the dialog to a much less politically delicate matter.
…Censorship is one thing we’ve come to count on from Chinese language mannequin builders and DeepSeek’s newest mannequin is not any exception.”
Nonetheless, as Perplexity AI’s CEO Aravind Srinivas confirmed, the mannequin could be uncensored. Opposite to some commenters on the LinkedIn dialogue, a self-hosted mannequin doesn’t telephone residence again to China, all the pieces is contained throughout the native surroundings.
Featured Picture by Shutterstock/gguy