Perplexity AI Deploys Chinese DeepSeek AI Model via @sejournal, @martinibuster

Perplexity AI has integrated the caller Chinese DeepSeek AI exemplary into their offerings, allowing their Pro level users to usage DeepSeek for their Perplexity AI research. Some successful the nationalist reacted negatively to the news.

Perplexity AI

Perplexity AI is San Francisco based AI hunt motor that offers a antithetic mode to hunt for accusation by leveraging web contented and ample connection models. There is besides a Pro Search mentation that allows unlimited record uploads, tin make images and offers a prime betwixt aggregate fashionable AI models similar OpenAI o1 and Anthropic’s Claude 3.5.

Now it’s offering DeepSeek R1 arsenic 1 of the disposable choices for Pro Users. The announcement was met with misconceptions astir what was being offered, including unfounded accusations that Perplexity DeepSeek information would beryllium accessible to the Chinese communist authorities and that the hunt results would beryllium censored.

Aravind Srinivas, Cofounder and CEO of Perplexity, commented connected LinkedIn astir the controversy:

“All DeepSeek usage successful Perplexity is done models hosted successful information centers successful the USA and Europe. DeepSeek is *open-source*. None of your information goes to China.”

The CEO besides took to X (formerly Twitter) to reassure users that the exemplary they are utilizing is not censored, posting a screenshot of an uncensored effect to show that the mentation of DeepSeek R1 successful usage astatine Perplexity is not censored

Screenshot Of Uncensored Perplexity AI DeepSeek R1

Is DeepSeek Self-Hosted Censored?

Anyone tin download the DeepSeek AI exemplary and usage it locally but the exemplary as-is volition beryllium censored since it’s lone bully arsenic the information it was trained on. The Register downloaded and tested aggregate models of DeepSeek and concluded that it is so censored:

“Is it censored?
Oh yeah. It is. Like galore Chinese models we’ve travel across, the DeepSeek R1 has been censored to forestall disapproval and embarrassment of the Chinese Communist Party.

Ask R1 astir delicate topics specified arsenic the 1989 Tiananmen Square massacre and we recovered it would outright garbage to entertain the question and effort to redirect the speech to a little politically delicate topic.

…Censorship is thing we’ve travel to expect from Chinese exemplary builders and DeepSeek’s latest exemplary is nary exception.”

However, arsenic Perplexity AI’s CEO Aravind Srinivas showed, the exemplary tin beryllium uncensored. Contrary to immoderate commenters connected the LinkedIn discussion, a self-hosted exemplary does not telephone location backmost to China, everything is contained wrong the section environment.

Featured Image by Shutterstock/gguy

SEJ STAFF Roger Montti Owner - Martinibuster.com astatine Martinibuster.com

I person 25 years hands-on acquisition successful SEO, evolving on with the hunt engines by keeping up with the latest ...